1 Git User's Manual (for version 1.5.1 or newer)
2 ______________________________________________
5 Git is a fast distributed revision control system.
7 This manual is designed to be readable by someone with basic unix
8 command-line skills, but no previous knowledge of git.
10 <<repositories-and-branches>> and <<exploring-git-history>> explain how
11 to fetch and study a project using git--read these chapters to learn how
12 to build and test a particular version of a software project, search for
13 regressions, and so on.
15 People needing to do actual development will also want to read
16 <<Developing-with-git>> and <<sharing-development>>.
18 Further chapters cover more specialized topics.
20 Comprehensive reference documentation is available through the man
21 pages. For a command such as "git clone", just use
23 ------------------------------------------------
25 ------------------------------------------------
27 See also <<git-quick-start>> for a brief overview of git commands,
28 without any explanation.
30 Also, see <<todo>> for ways that you can help make this manual more
34 [[repositories-and-branches]]
35 Repositories and Branches
36 =========================
38 [[how-to-get-a-git-repository]]
39 How to get a git repository
40 ---------------------------
42 It will be useful to have a git repository to experiment with as you
45 The best way to get one is by using the gitlink:git-clone[1] command
46 to download a copy of an existing repository for a project that you
47 are interested in. If you don't already have a project in mind, here
48 are some interesting examples:
50 ------------------------------------------------
51 # git itself (approx. 10MB download):
52 $ git clone git://git.kernel.org/pub/scm/git/git.git
53 # the linux kernel (approx. 150MB download):
54 $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
55 ------------------------------------------------
57 The initial clone may be time-consuming for a large project, but you
58 will only need to clone once.
60 The clone command creates a new directory named after the project
61 ("git" or "linux-2.6" in the examples above). After you cd into this
62 directory, you will see that it contains a copy of the project files,
63 together with a special top-level directory named ".git", which
64 contains all the information about the history of the project.
66 In most of the following, examples will be taken from one of the two
70 How to check out a different version of a project
71 -------------------------------------------------
73 Git is best thought of as a tool for storing the history of a
74 collection of files. It stores the history as a compressed
75 collection of interrelated snapshots (versions) of the project's
78 A single git repository may contain multiple branches. It keeps track
79 of them by keeping a list of <<def_head,heads>> which reference the
80 latest version on each branch; the gitlink:git-branch[1] command shows
81 you the list of branch heads:
83 ------------------------------------------------
86 ------------------------------------------------
88 A freshly cloned repository contains a single branch head, by default
89 named "master", with the working directory initialized to the state of
90 the project referred to by that branch head.
92 Most projects also use <<def_tag,tags>>. Tags, like heads, are
93 references into the project's history, and can be listed using the
94 gitlink:git-tag[1] command:
96 ------------------------------------------------
108 ------------------------------------------------
110 Tags are expected to always point at the same version of a project,
111 while heads are expected to advance as development progresses.
113 Create a new branch head pointing to one of these versions and check it
114 out using gitlink:git-checkout[1]:
116 ------------------------------------------------
117 $ git checkout -b new v2.6.13
118 ------------------------------------------------
120 The working directory then reflects the contents that the project had
121 when it was tagged v2.6.13, and gitlink:git-branch[1] shows two
122 branches, with an asterisk marking the currently checked-out branch:
124 ------------------------------------------------
128 ------------------------------------------------
130 If you decide that you'd rather see version 2.6.17, you can modify
131 the current branch to point at v2.6.17 instead, with
133 ------------------------------------------------
134 $ git reset --hard v2.6.17
135 ------------------------------------------------
137 Note that if the current branch head was your only reference to a
138 particular point in history, then resetting that branch may leave you
139 with no way to find the history it used to point to; so use this command
142 [[understanding-commits]]
143 Understanding History: Commits
144 ------------------------------
146 Every change in the history of a project is represented by a commit.
147 The gitlink:git-show[1] command shows the most recent commit on the
150 ------------------------------------------------
152 commit 2b5f6dcce5bf94b9b119e9ed8d537098ec61c3d2
153 Author: Jamal Hadi Salim <hadi@cyberus.ca>
154 Date: Sat Dec 2 22:22:25 2006 -0800
156 [XFRM]: Fix aevent structuring to be more complete.
158 aevents can not uniquely identify an SA. We break the ABI with this
159 patch, but consensus is that since it is not yet utilized by any
160 (known) application then it is fine (better do it now than later).
162 Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
163 Signed-off-by: David S. Miller <davem@davemloft.net>
165 diff --git a/Documentation/networking/xfrm_sync.txt b/Documentation/networking/xfrm_sync.txt
166 index 8be626f..d7aac9d 100644
167 --- a/Documentation/networking/xfrm_sync.txt
168 +++ b/Documentation/networking/xfrm_sync.txt
169 @@ -47,10 +47,13 @@ aevent_id structure looks like:
171 struct xfrm_aevent_id {
172 struct xfrm_usersa_id sa_id;
173 + xfrm_address_t saddr;
178 ------------------------------------------------
180 As you can see, a commit shows who made the latest change, what they
183 Every commit has a 40-hexdigit id, sometimes called the "object name" or the
184 "SHA1 id", shown on the first line of the "git show" output. You can usually
185 refer to a commit by a shorter name, such as a tag or a branch name, but this
186 longer name can also be useful. Most importantly, it is a globally unique
187 name for this commit: so if you tell somebody else the object name (for
188 example in email), then you are guaranteed that name will refer to the same
189 commit in their repository that it does in yours (assuming their repository
190 has that commit at all). Since the object name is computed as a hash over the
191 contents of the commit, you are guaranteed that the commit can never change
192 without its name also changing.
194 In fact, in <<git-internals>> we shall see that everything stored in git
195 history, including file data and directory contents, is stored in an object
196 with a name that is a hash of its contents.
198 [[understanding-reachability]]
199 Understanding history: commits, parents, and reachability
200 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
202 Every commit (except the very first commit in a project) also has a
203 parent commit which shows what happened before this commit.
204 Following the chain of parents will eventually take you back to the
205 beginning of the project.
207 However, the commits do not form a simple list; git allows lines of
208 development to diverge and then reconverge, and the point where two
209 lines of development reconverge is called a "merge". The commit
210 representing a merge can therefore have more than one parent, with
211 each parent representing the most recent commit on one of the lines
212 of development leading to that point.
214 The best way to see how this works is using the gitlink:gitk[1]
215 command; running gitk now on a git repository and looking for merge
216 commits will help understand how the git organizes history.
218 In the following, we say that commit X is "reachable" from commit Y
219 if commit X is an ancestor of commit Y. Equivalently, you could say
220 that Y is a descendent of X, or that there is a chain of parents
221 leading from commit Y to commit X.
224 Understanding history: History diagrams
225 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
227 We will sometimes represent git history using diagrams like the one
228 below. Commits are shown as "o", and the links between them with
229 lines drawn with - / and \. Time goes left to right:
232 ................................................
238 ................................................
240 If we need to talk about a particular commit, the character "o" may
241 be replaced with another letter or number.
244 Understanding history: What is a branch?
245 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
247 When we need to be precise, we will use the word "branch" to mean a line
248 of development, and "branch head" (or just "head") to mean a reference
249 to the most recent commit on a branch. In the example above, the branch
250 head named "A" is a pointer to one particular commit, but we refer to
251 the line of three commits leading up to that point as all being part of
254 However, when no confusion will result, we often just use the term
255 "branch" both for branches and for branch heads.
257 [[manipulating-branches]]
258 Manipulating branches
259 ---------------------
261 Creating, deleting, and modifying branches is quick and easy; here's
262 a summary of the commands:
266 git branch <branch>::
267 create a new branch named <branch>, referencing the same
268 point in history as the current branch
269 git branch <branch> <start-point>::
270 create a new branch named <branch>, referencing
271 <start-point>, which may be specified any way you like,
272 including using a branch name or a tag name
273 git branch -d <branch>::
274 delete the branch <branch>; if the branch you are deleting
275 points to a commit which is not reachable from the current
276 branch, this command will fail with a warning.
277 git branch -D <branch>::
278 even if the branch points to a commit not reachable
279 from the current branch, you may know that that commit
280 is still reachable from some other branch or tag. In that
281 case it is safe to use this command to force git to delete
283 git checkout <branch>::
284 make the current branch <branch>, updating the working
285 directory to reflect the version referenced by <branch>
286 git checkout -b <new> <start-point>::
287 create a new branch <new> referencing <start-point>, and
290 The special symbol "HEAD" can always be used to refer to the current
291 branch. In fact, git uses a file named "HEAD" in the .git directory to
292 remember which branch is current:
294 ------------------------------------------------
296 ref: refs/heads/master
297 ------------------------------------------------
300 Examining an old version without creating a new branch
301 ------------------------------------------------------
303 The git-checkout command normally expects a branch head, but will also
304 accept an arbitrary commit; for example, you can check out the commit
307 ------------------------------------------------
308 $ git checkout v2.6.17
309 Note: moving to "v2.6.17" which isn't a local branch
310 If you want to create a new branch from this checkout, you may do so
311 (now or later) by using -b with the checkout command again. Example:
312 git checkout -b <new_branch_name>
313 HEAD is now at 427abfa... Linux v2.6.17
314 ------------------------------------------------
316 The HEAD then refers to the SHA1 of the commit instead of to a branch,
317 and git branch shows that you are no longer on a branch:
319 ------------------------------------------------
321 427abfa28afedffadfca9dd8b067eb6d36bac53f
325 ------------------------------------------------
327 In this case we say that the HEAD is "detached".
329 This is an easy way to check out a particular version without having to
330 make up a name for the new branch. You can still create a new branch
331 (or tag) for this version later if you decide to.
333 [[examining-remote-branches]]
334 Examining branches from a remote repository
335 -------------------------------------------
337 The "master" branch that was created at the time you cloned is a copy
338 of the HEAD in the repository that you cloned from. That repository
339 may also have had other branches, though, and your local repository
340 keeps branches which track each of those remote branches, which you
341 can view using the "-r" option to gitlink:git-branch[1]:
343 ------------------------------------------------
353 ------------------------------------------------
355 You cannot check out these remote-tracking branches, but you can
356 examine them on a branch of your own, just as you would a tag:
358 ------------------------------------------------
359 $ git checkout -b my-todo-copy origin/todo
360 ------------------------------------------------
362 Note that the name "origin" is just the name that git uses by default
363 to refer to the repository that you cloned from.
365 [[how-git-stores-references]]
366 Naming branches, tags, and other references
367 -------------------------------------------
369 Branches, remote-tracking branches, and tags are all references to
370 commits. All references are named with a slash-separated path name
371 starting with "refs"; the names we've been using so far are actually
374 - The branch "test" is short for "refs/heads/test".
375 - The tag "v2.6.18" is short for "refs/tags/v2.6.18".
376 - "origin/master" is short for "refs/remotes/origin/master".
378 The full name is occasionally useful if, for example, there ever
379 exists a tag and a branch with the same name.
381 As another useful shortcut, the "HEAD" of a repository can be referred
382 to just using the name of that repository. So, for example, "origin"
383 is usually a shortcut for the HEAD branch in the repository "origin".
385 For the complete list of paths which git checks for references, and
386 the order it uses to decide which to choose when there are multiple
387 references with the same shorthand name, see the "SPECIFYING
388 REVISIONS" section of gitlink:git-rev-parse[1].
390 [[Updating-a-repository-with-git-fetch]]
391 Updating a repository with git fetch
392 ------------------------------------
394 Eventually the developer cloned from will do additional work in her
395 repository, creating new commits and advancing the branches to point
398 The command "git fetch", with no arguments, will update all of the
399 remote-tracking branches to the latest version found in her
400 repository. It will not touch any of your own branches--not even the
401 "master" branch that was created for you on clone.
403 [[fetching-branches]]
404 Fetching branches from other repositories
405 -----------------------------------------
407 You can also track branches from repositories other than the one you
408 cloned from, using gitlink:git-remote[1]:
410 -------------------------------------------------
411 $ git remote add linux-nfs git://linux-nfs.org/pub/nfs-2.6.git
412 $ git fetch linux-nfs
413 * refs/remotes/linux-nfs/master: storing branch 'master' ...
415 -------------------------------------------------
417 New remote-tracking branches will be stored under the shorthand name
418 that you gave "git remote add", in this case linux-nfs:
420 -------------------------------------------------
424 -------------------------------------------------
426 If you run "git fetch <remote>" later, the tracking branches for the
427 named <remote> will be updated.
429 If you examine the file .git/config, you will see that git has added
432 -------------------------------------------------
436 url = git://linux-nfs.org/pub/nfs-2.6.git
437 fetch = +refs/heads/*:refs/remotes/linux-nfs/*
439 -------------------------------------------------
441 This is what causes git to track the remote's branches; you may modify
442 or delete these configuration options by editing .git/config with a
443 text editor. (See the "CONFIGURATION FILE" section of
444 gitlink:git-config[1] for details.)
446 [[exploring-git-history]]
447 Exploring git history
448 =====================
450 Git is best thought of as a tool for storing the history of a
451 collection of files. It does this by storing compressed snapshots of
452 the contents of a file heirarchy, together with "commits" which show
453 the relationships between these snapshots.
455 Git provides extremely flexible and fast tools for exploring the
456 history of a project.
458 We start with one specialized tool that is useful for finding the
459 commit that introduced a bug into a project.
462 How to use bisect to find a regression
463 --------------------------------------
465 Suppose version 2.6.18 of your project worked, but the version at
466 "master" crashes. Sometimes the best way to find the cause of such a
467 regression is to perform a brute-force search through the project's
468 history to find the particular commit that caused the problem. The
469 gitlink:git-bisect[1] command can help you do this:
471 -------------------------------------------------
473 $ git bisect good v2.6.18
474 $ git bisect bad master
475 Bisecting: 3537 revisions left to test after this
476 [65934a9a028b88e83e2b0f8b36618fe503349f8e] BLOCK: Make USB storage depend on SCSI rather than selecting it [try #6]
477 -------------------------------------------------
479 If you run "git branch" at this point, you'll see that git has
480 temporarily moved you to a new branch named "bisect". This branch
481 points to a commit (with commit id 65934...) that is reachable from
482 v2.6.19 but not from v2.6.18. Compile and test it, and see whether
483 it crashes. Assume it does crash. Then:
485 -------------------------------------------------
487 Bisecting: 1769 revisions left to test after this
488 [7eff82c8b1511017ae605f0c99ac275a7e21b867] i2c-core: Drop useless bitmaskings
489 -------------------------------------------------
491 checks out an older version. Continue like this, telling git at each
492 stage whether the version it gives you is good or bad, and notice
493 that the number of revisions left to test is cut approximately in
496 After about 13 tests (in this case), it will output the commit id of
497 the guilty commit. You can then examine the commit with
498 gitlink:git-show[1], find out who wrote it, and mail them your bug
499 report with the commit id. Finally, run
501 -------------------------------------------------
503 -------------------------------------------------
505 to return you to the branch you were on before and delete the
506 temporary "bisect" branch.
508 Note that the version which git-bisect checks out for you at each
509 point is just a suggestion, and you're free to try a different
510 version if you think it would be a good idea. For example,
511 occasionally you may land on a commit that broke something unrelated;
514 -------------------------------------------------
515 $ git bisect visualize
516 -------------------------------------------------
518 which will run gitk and label the commit it chose with a marker that
519 says "bisect". Chose a safe-looking commit nearby, note its commit
520 id, and check it out with:
522 -------------------------------------------------
523 $ git reset --hard fb47ddb2db...
524 -------------------------------------------------
526 then test, run "bisect good" or "bisect bad" as appropriate, and
533 We have seen several ways of naming commits already:
535 - 40-hexdigit object name
536 - branch name: refers to the commit at the head of the given
538 - tag name: refers to the commit pointed to by the given tag
539 (we've seen branches and tags are special cases of
540 <<how-git-stores-references,references>>).
541 - HEAD: refers to the head of the current branch
543 There are many more; see the "SPECIFYING REVISIONS" section of the
544 gitlink:git-rev-parse[1] man page for the complete list of ways to
545 name revisions. Some examples:
547 -------------------------------------------------
548 $ git show fb47ddb2 # the first few characters of the object name
549 # are usually enough to specify it uniquely
550 $ git show HEAD^ # the parent of the HEAD commit
551 $ git show HEAD^^ # the grandparent
552 $ git show HEAD~4 # the great-great-grandparent
553 -------------------------------------------------
555 Recall that merge commits may have more than one parent; by default,
556 ^ and ~ follow the first parent listed in the commit, but you can
559 -------------------------------------------------
560 $ git show HEAD^1 # show the first parent of HEAD
561 $ git show HEAD^2 # show the second parent of HEAD
562 -------------------------------------------------
564 In addition to HEAD, there are several other special names for
567 Merges (to be discussed later), as well as operations such as
568 git-reset, which change the currently checked-out commit, generally
569 set ORIG_HEAD to the value HEAD had before the current operation.
571 The git-fetch operation always stores the head of the last fetched
572 branch in FETCH_HEAD. For example, if you run git fetch without
573 specifying a local branch as the target of the operation
575 -------------------------------------------------
576 $ git fetch git://example.com/proj.git theirbranch
577 -------------------------------------------------
579 the fetched commits will still be available from FETCH_HEAD.
581 When we discuss merges we'll also see the special name MERGE_HEAD,
582 which refers to the other branch that we're merging in to the current
585 The gitlink:git-rev-parse[1] command is a low-level command that is
586 occasionally useful for translating some name for a commit to the object
587 name for that commit:
589 -------------------------------------------------
590 $ git rev-parse origin
591 e05db0fd4f31dde7005f075a84f96b360d05984b
592 -------------------------------------------------
598 We can also create a tag to refer to a particular commit; after
601 -------------------------------------------------
602 $ git tag stable-1 1b2e1d63ff
603 -------------------------------------------------
605 You can use stable-1 to refer to the commit 1b2e1d63ff.
607 This creates a "lightweight" tag. If you would also like to include a
608 comment with the tag, and possibly sign it cryptographically, then you
609 should create a tag object instead; see the gitlink:git-tag[1] man page
612 [[browsing-revisions]]
616 The gitlink:git-log[1] command can show lists of commits. On its
617 own, it shows all commits reachable from the parent commit; but you
618 can also make more specific requests:
620 -------------------------------------------------
621 $ git log v2.5.. # commits since (not reachable from) v2.5
622 $ git log test..master # commits reachable from master but not test
623 $ git log master..test # ...reachable from test but not master
624 $ git log master...test # ...reachable from either test or master,
626 $ git log --since="2 weeks ago" # commits from the last 2 weeks
627 $ git log Makefile # commits which modify Makefile
628 $ git log fs/ # ... which modify any file under fs/
629 $ git log -S'foo()' # commits which add or remove any file data
630 # matching the string 'foo()'
631 -------------------------------------------------
633 And of course you can combine all of these; the following finds
634 commits since v2.5 which touch the Makefile or any file under fs:
636 -------------------------------------------------
637 $ git log v2.5.. Makefile fs/
638 -------------------------------------------------
640 You can also ask git log to show patches:
642 -------------------------------------------------
644 -------------------------------------------------
646 See the "--pretty" option in the gitlink:git-log[1] man page for more
649 Note that git log starts with the most recent commit and works
650 backwards through the parents; however, since git history can contain
651 multiple independent lines of development, the particular order that
652 commits are listed in may be somewhat arbitrary.
658 You can generate diffs between any two versions using
661 -------------------------------------------------
662 $ git diff master..test
663 -------------------------------------------------
665 Sometimes what you want instead is a set of patches:
667 -------------------------------------------------
668 $ git format-patch master..test
669 -------------------------------------------------
671 will generate a file with a patch for each commit reachable from test
672 but not from master. Note that if master also has commits which are
673 not reachable from test, then the combined result of these patches
674 will not be the same as the diff produced by the git-diff example.
676 [[viewing-old-file-versions]]
677 Viewing old file versions
678 -------------------------
680 You can always view an old version of a file by just checking out the
681 correct revision first. But sometimes it is more convenient to be
682 able to view an old version of a single file without checking
683 anything out; this command does that:
685 -------------------------------------------------
686 $ git show v2.5:fs/locks.c
687 -------------------------------------------------
689 Before the colon may be anything that names a commit, and after it
690 may be any path to a file tracked by git.
696 [[counting-commits-on-a-branch]]
697 Counting the number of commits on a branch
698 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
700 Suppose you want to know how many commits you've made on "mybranch"
701 since it diverged from "origin":
703 -------------------------------------------------
704 $ git log --pretty=oneline origin..mybranch | wc -l
705 -------------------------------------------------
707 Alternatively, you may often see this sort of thing done with the
708 lower-level command gitlink:git-rev-list[1], which just lists the SHA1's
709 of all the given commits:
711 -------------------------------------------------
712 $ git rev-list origin..mybranch | wc -l
713 -------------------------------------------------
715 [[checking-for-equal-branches]]
716 Check whether two branches point at the same history
717 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
719 Suppose you want to check whether two branches point at the same point
722 -------------------------------------------------
723 $ git diff origin..master
724 -------------------------------------------------
726 will tell you whether the contents of the project are the same at the
727 two branches; in theory, however, it's possible that the same project
728 contents could have been arrived at by two different historical
729 routes. You could compare the object names:
731 -------------------------------------------------
732 $ git rev-list origin
733 e05db0fd4f31dde7005f075a84f96b360d05984b
734 $ git rev-list master
735 e05db0fd4f31dde7005f075a84f96b360d05984b
736 -------------------------------------------------
738 Or you could recall that the ... operator selects all commits
739 contained reachable from either one reference or the other but not
742 -------------------------------------------------
743 $ git log origin...master
744 -------------------------------------------------
746 will return no commits when the two branches are equal.
748 [[finding-tagged-descendants]]
749 Find first tagged version including a given fix
750 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
752 Suppose you know that the commit e05db0fd fixed a certain problem.
753 You'd like to find the earliest tagged release that contains that
756 Of course, there may be more than one answer--if the history branched
757 after commit e05db0fd, then there could be multiple "earliest" tagged
760 You could just visually inspect the commits since e05db0fd:
762 -------------------------------------------------
764 -------------------------------------------------
766 Or you can use gitlink:git-name-rev[1], which will give the commit a
767 name based on any tag it finds pointing to one of the commit's
770 -------------------------------------------------
771 $ git name-rev --tags e05db0fd
772 e05db0fd tags/v1.5.0-rc1^0~23
773 -------------------------------------------------
775 The gitlink:git-describe[1] command does the opposite, naming the
776 revision using a tag on which the given commit is based:
778 -------------------------------------------------
779 $ git describe e05db0fd
780 v1.5.0-rc0-260-ge05db0f
781 -------------------------------------------------
783 but that may sometimes help you guess which tags might come after the
786 If you just want to verify whether a given tagged version contains a
787 given commit, you could use gitlink:git-merge-base[1]:
789 -------------------------------------------------
790 $ git merge-base e05db0fd v1.5.0-rc1
791 e05db0fd4f31dde7005f075a84f96b360d05984b
792 -------------------------------------------------
794 The merge-base command finds a common ancestor of the given commits,
795 and always returns one or the other in the case where one is a
796 descendant of the other; so the above output shows that e05db0fd
797 actually is an ancestor of v1.5.0-rc1.
799 Alternatively, note that
801 -------------------------------------------------
802 $ git log v1.5.0-rc1..e05db0fd
803 -------------------------------------------------
805 will produce empty output if and only if v1.5.0-rc1 includes e05db0fd,
806 because it outputs only commits that are not reachable from v1.5.0-rc1.
808 As yet another alternative, the gitlink:git-show-branch[1] command lists
809 the commits reachable from its arguments with a display on the left-hand
810 side that indicates which arguments that commit is reachable from. So,
811 you can run something like
813 -------------------------------------------------
814 $ git show-branch e05db0fd v1.5.0-rc0 v1.5.0-rc1 v1.5.0-rc2
815 ! [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
817 ! [v1.5.0-rc0] GIT v1.5.0 preview
818 ! [v1.5.0-rc1] GIT v1.5.0-rc1
819 ! [v1.5.0-rc2] GIT v1.5.0-rc2
821 -------------------------------------------------
823 then search for a line that looks like
825 -------------------------------------------------
826 + ++ [e05db0fd] Fix warnings in sha1_file.c - use C99 printf format if
828 -------------------------------------------------
830 Which shows that e05db0fd is reachable from itself, from v1.5.0-rc1, and
831 from v1.5.0-rc2, but not from v1.5.0-rc0.
834 Creating a changelog and tarball for a software release
835 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
837 The gitlink:git-archive[1] command can create a tar or zip archive from
838 any version of a project; for example:
840 -------------------------------------------------
841 $ git archive --format=tar --prefix=project/ HEAD | gzip >latest.tar.gz
842 -------------------------------------------------
844 will use HEAD to produce a tar archive in which each filename is
845 preceded by "prefix/".
847 If you're releasing a new version of a software project, you may want
848 to simultaneously make a changelog to include in the release
851 Linus Torvalds, for example, makes new kernel releases by tagging them,
854 -------------------------------------------------
855 $ release-script 2.6.12 2.6.13-rc6 2.6.13-rc7
856 -------------------------------------------------
858 where release-script is a shell script that looks like:
860 -------------------------------------------------
865 echo "# git tag v$new"
866 echo "git archive --prefix=linux-$new/ v$new | gzip -9 > ../linux-$new.tar.gz"
867 echo "git diff v$stable v$new | gzip -9 > ../patch-$new.gz"
868 echo "git log --no-merges v$new ^v$last > ../ChangeLog-$new"
869 echo "git shortlog --no-merges v$new ^v$last > ../ShortLog"
870 echo "git diff --stat --summary -M v$last v$new > ../diffstat-$new"
871 -------------------------------------------------
873 and then he just cut-and-pastes the output commands after verifying that
876 [[Developing-with-git]]
880 [[telling-git-your-name]]
881 Telling git your name
882 ---------------------
884 Before creating any commits, you should introduce yourself to git. The
885 easiest way to do so is to make sure the following lines appear in a
886 file named .gitconfig in your home directory:
888 ------------------------------------------------
890 name = Your Name Comes Here
891 email = you@yourdomain.example.com
892 ------------------------------------------------
894 (See the "CONFIGURATION FILE" section of gitlink:git-config[1] for
895 details on the configuration file.)
898 [[creating-a-new-repository]]
899 Creating a new repository
900 -------------------------
902 Creating a new repository from scratch is very easy:
904 -------------------------------------------------
908 -------------------------------------------------
910 If you have some initial content (say, a tarball):
912 -------------------------------------------------
913 $ tar -xzvf project.tar.gz
916 $ git add . # include everything below ./ in the first commit:
918 -------------------------------------------------
920 [[how-to-make-a-commit]]
924 Creating a new commit takes three steps:
926 1. Making some changes to the working directory using your
928 2. Telling git about your changes.
929 3. Creating the commit using the content you told git about
932 In practice, you can interleave and repeat steps 1 and 2 as many
933 times as you want: in order to keep track of what you want committed
934 at step 3, git maintains a snapshot of the tree's contents in a
935 special staging area called "the index."
937 At the beginning, the content of the index will be identical to
938 that of the HEAD. The command "git diff --cached", which shows
939 the difference between the HEAD and the index, should therefore
940 produce no output at that point.
942 Modifying the index is easy:
944 To update the index with the new contents of a modified file, use
946 -------------------------------------------------
947 $ git add path/to/file
948 -------------------------------------------------
950 To add the contents of a new file to the index, use
952 -------------------------------------------------
953 $ git add path/to/file
954 -------------------------------------------------
956 To remove a file from the index and from the working tree,
958 -------------------------------------------------
959 $ git rm path/to/file
960 -------------------------------------------------
962 After each step you can verify that
964 -------------------------------------------------
966 -------------------------------------------------
968 always shows the difference between the HEAD and the index file--this
969 is what you'd commit if you created the commit now--and that
971 -------------------------------------------------
973 -------------------------------------------------
975 shows the difference between the working tree and the index file.
977 Note that "git add" always adds just the current contents of a file
978 to the index; further changes to the same file will be ignored unless
979 you run git-add on the file again.
981 When you're ready, just run
983 -------------------------------------------------
985 -------------------------------------------------
987 and git will prompt you for a commit message and then create the new
988 commit. Check to make sure it looks like what you expected with
990 -------------------------------------------------
992 -------------------------------------------------
994 As a special shortcut,
996 -------------------------------------------------
998 -------------------------------------------------
1000 will update the index with any files that you've modified or removed
1001 and create a commit, all in one step.
1003 A number of commands are useful for keeping track of what you're
1006 -------------------------------------------------
1007 $ git diff --cached # difference between HEAD and the index; what
1008 # would be commited if you ran "commit" now.
1009 $ git diff # difference between the index file and your
1010 # working directory; changes that would not
1011 # be included if you ran "commit" now.
1012 $ git diff HEAD # difference between HEAD and working tree; what
1013 # would be committed if you ran "commit -a" now.
1014 $ git status # a brief per-file summary of the above.
1015 -------------------------------------------------
1017 [[creating-good-commit-messages]]
1018 Creating good commit messages
1019 -----------------------------
1021 Though not required, it's a good idea to begin the commit message
1022 with a single short (less than 50 character) line summarizing the
1023 change, followed by a blank line and then a more thorough
1024 description. Tools that turn commits into email, for example, use
1025 the first line on the Subject line and the rest of the commit in the
1032 You can rejoin two diverging branches of development using
1033 gitlink:git-merge[1]:
1035 -------------------------------------------------
1036 $ git merge branchname
1037 -------------------------------------------------
1039 merges the development in the branch "branchname" into the current
1040 branch. If there are conflicts--for example, if the same file is
1041 modified in two different ways in the remote branch and the local
1042 branch--then you are warned; the output may look something like this:
1044 -------------------------------------------------
1047 Auto-merged file.txt
1048 CONFLICT (content): Merge conflict in file.txt
1049 Automatic merge failed; fix conflicts and then commit the result.
1050 -------------------------------------------------
1052 Conflict markers are left in the problematic files, and after
1053 you resolve the conflicts manually, you can update the index
1054 with the contents and run git commit, as you normally would when
1055 creating a new file.
1057 If you examine the resulting commit using gitk, you will see that it
1058 has two parents, one pointing to the top of the current branch, and
1059 one to the top of the other branch.
1061 [[resolving-a-merge]]
1065 When a merge isn't resolved automatically, git leaves the index and
1066 the working tree in a special state that gives you all the
1067 information you need to help resolve the merge.
1069 Files with conflicts are marked specially in the index, so until you
1070 resolve the problem and update the index, gitlink:git-commit[1] will
1073 -------------------------------------------------
1075 file.txt: needs merge
1076 -------------------------------------------------
1078 Also, gitlink:git-status[1] will list those files as "unmerged", and the
1079 files with conflicts will have conflict markers added, like this:
1081 -------------------------------------------------
1082 <<<<<<< HEAD:file.txt
1086 >>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1087 -------------------------------------------------
1089 All you need to do is edit the files to resolve the conflicts, and then
1091 -------------------------------------------------
1094 -------------------------------------------------
1096 Note that the commit message will already be filled in for you with
1097 some information about the merge. Normally you can just use this
1098 default message unchanged, but you may add additional commentary of
1099 your own if desired.
1101 The above is all you need to know to resolve a simple merge. But git
1102 also provides more information to help resolve conflicts:
1104 [[conflict-resolution]]
1105 Getting conflict-resolution help during a merge
1106 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1108 All of the changes that git was able to merge automatically are
1109 already added to the index file, so gitlink:git-diff[1] shows only
1110 the conflicts. It uses an unusual syntax:
1112 -------------------------------------------------
1115 index 802992c,2b60207..0000000
1118 @@@ -1,1 -1,1 +1,5 @@@
1119 ++<<<<<<< HEAD:file.txt
1123 ++>>>>>>> 77976da35a11db4580b80ae27e8d65caf5208086:file.txt
1124 -------------------------------------------------
1126 Recall that the commit which will be commited after we resolve this
1127 conflict will have two parents instead of the usual one: one parent
1128 will be HEAD, the tip of the current branch; the other will be the
1129 tip of the other branch, which is stored temporarily in MERGE_HEAD.
1131 During the merge, the index holds three versions of each file. Each of
1132 these three "file stages" represents a different version of the file:
1134 -------------------------------------------------
1135 $ git show :1:file.txt # the file in a common ancestor of both branches
1136 $ git show :2:file.txt # the version from HEAD, but including any
1137 # nonconflicting changes from MERGE_HEAD
1138 $ git show :3:file.txt # the version from MERGE_HEAD, but including any
1139 # nonconflicting changes from HEAD.
1140 -------------------------------------------------
1142 Since the stage 2 and stage 3 versions have already been updated with
1143 nonconflicting changes, the only remaining differences between them are
1144 the important ones; thus gitlink:git-diff[1] can use the information in
1145 the index to show only those conflicts.
1147 The diff above shows the differences between the working-tree version of
1148 file.txt and the stage 2 and stage 3 versions. So instead of preceding
1149 each line by a single "+" or "-", it now uses two columns: the first
1150 column is used for differences between the first parent and the working
1151 directory copy, and the second for differences between the second parent
1152 and the working directory copy. (See the "COMBINED DIFF FORMAT" section
1153 of gitlink:git-diff-files[1] for a details of the format.)
1155 After resolving the conflict in the obvious way (but before updating the
1156 index), the diff will look like:
1158 -------------------------------------------------
1161 index 802992c,2b60207..0000000
1164 @@@ -1,1 -1,1 +1,1 @@@
1168 -------------------------------------------------
1170 This shows that our resolved version deleted "Hello world" from the
1171 first parent, deleted "Goodbye" from the second parent, and added
1172 "Goodbye world", which was previously absent from both.
1174 Some special diff options allow diffing the working directory against
1175 any of these stages:
1177 -------------------------------------------------
1178 $ git diff -1 file.txt # diff against stage 1
1179 $ git diff --base file.txt # same as the above
1180 $ git diff -2 file.txt # diff against stage 2
1181 $ git diff --ours file.txt # same as the above
1182 $ git diff -3 file.txt # diff against stage 3
1183 $ git diff --theirs file.txt # same as the above.
1184 -------------------------------------------------
1186 The gitlink:git-log[1] and gitk[1] commands also provide special help
1189 -------------------------------------------------
1192 -------------------------------------------------
1194 These will display all commits which exist only on HEAD or on
1195 MERGE_HEAD, and which touch an unmerged file.
1197 You may also use gitlink:git-mergetool[1], which lets you merge the
1198 unmerged files using external tools such as emacs or kdiff3.
1200 Each time you resolve the conflicts in a file and update the index:
1202 -------------------------------------------------
1204 -------------------------------------------------
1206 the different stages of that file will be "collapsed", after which
1207 git-diff will (by default) no longer show diffs for that file.
1213 If you get stuck and decide to just give up and throw the whole mess
1214 away, you can always return to the pre-merge state with
1216 -------------------------------------------------
1217 $ git reset --hard HEAD
1218 -------------------------------------------------
1220 Or, if you've already commited the merge that you want to throw away,
1222 -------------------------------------------------
1223 $ git reset --hard ORIG_HEAD
1224 -------------------------------------------------
1226 However, this last command can be dangerous in some cases--never
1227 throw away a commit you have already committed if that commit may
1228 itself have been merged into another branch, as doing so may confuse
1235 There is one special case not mentioned above, which is treated
1236 differently. Normally, a merge results in a merge commit, with two
1237 parents, one pointing at each of the two lines of development that
1240 However, if the current branch is a descendant of the other--so every
1241 commit present in the one is already contained in the other--then git
1242 just performs a "fast forward"; the head of the current branch is moved
1243 forward to point at the head of the merged-in branch, without any new
1244 commits being created.
1250 If you've messed up the working tree, but haven't yet committed your
1251 mistake, you can return the entire working tree to the last committed
1254 -------------------------------------------------
1255 $ git reset --hard HEAD
1256 -------------------------------------------------
1258 If you make a commit that you later wish you hadn't, there are two
1259 fundamentally different ways to fix the problem:
1261 1. You can create a new commit that undoes whatever was done
1262 by the previous commit. This is the correct thing if your
1263 mistake has already been made public.
1265 2. You can go back and modify the old commit. You should
1266 never do this if you have already made the history public;
1267 git does not normally expect the "history" of a project to
1268 change, and cannot correctly perform repeated merges from
1269 a branch that has had its history changed.
1271 [[reverting-a-commit]]
1272 Fixing a mistake with a new commit
1273 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1275 Creating a new commit that reverts an earlier change is very easy;
1276 just pass the gitlink:git-revert[1] command a reference to the bad
1277 commit; for example, to revert the most recent commit:
1279 -------------------------------------------------
1281 -------------------------------------------------
1283 This will create a new commit which undoes the change in HEAD. You
1284 will be given a chance to edit the commit message for the new commit.
1286 You can also revert an earlier change, for example, the next-to-last:
1288 -------------------------------------------------
1290 -------------------------------------------------
1292 In this case git will attempt to undo the old change while leaving
1293 intact any changes made since then. If more recent changes overlap
1294 with the changes to be reverted, then you will be asked to fix
1295 conflicts manually, just as in the case of <<resolving-a-merge,
1296 resolving a merge>>.
1298 [[fixing-a-mistake-by-editing-history]]
1299 Fixing a mistake by editing history
1300 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1302 If the problematic commit is the most recent commit, and you have not
1303 yet made that commit public, then you may just
1304 <<undoing-a-merge,destroy it using git-reset>>.
1307 can edit the working directory and update the index to fix your
1308 mistake, just as if you were going to <<how-to-make-a-commit,create a
1309 new commit>>, then run
1311 -------------------------------------------------
1312 $ git commit --amend
1313 -------------------------------------------------
1315 which will replace the old commit by a new commit incorporating your
1316 changes, giving you a chance to edit the old commit message first.
1318 Again, you should never do this to a commit that may already have
1319 been merged into another branch; use gitlink:git-revert[1] instead in
1322 It is also possible to edit commits further back in the history, but
1323 this is an advanced topic to be left for
1324 <<cleaning-up-history,another chapter>>.
1326 [[checkout-of-path]]
1327 Checking out an old version of a file
1328 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1330 In the process of undoing a previous bad change, you may find it
1331 useful to check out an older version of a particular file using
1332 gitlink:git-checkout[1]. We've used git checkout before to switch
1333 branches, but it has quite different behavior if it is given a path
1336 -------------------------------------------------
1337 $ git checkout HEAD^ path/to/file
1338 -------------------------------------------------
1340 replaces path/to/file by the contents it had in the commit HEAD^, and
1341 also updates the index to match. It does not change branches.
1343 If you just want to look at an old version of the file, without
1344 modifying the working directory, you can do that with
1345 gitlink:git-show[1]:
1347 -------------------------------------------------
1348 $ git show HEAD^:path/to/file
1349 -------------------------------------------------
1351 which will display the given version of the file.
1353 [[ensuring-good-performance]]
1354 Ensuring good performance
1355 -------------------------
1357 On large repositories, git depends on compression to keep the history
1358 information from taking up to much space on disk or in memory.
1360 This compression is not performed automatically. Therefore you
1361 should occasionally run gitlink:git-gc[1]:
1363 -------------------------------------------------
1365 -------------------------------------------------
1367 to recompress the archive. This can be very time-consuming, so
1368 you may prefer to run git-gc when you are not doing other work.
1371 [[ensuring-reliability]]
1372 Ensuring reliability
1373 --------------------
1375 [[checking-for-corruption]]
1376 Checking the repository for corruption
1377 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1379 The gitlink:git-fsck[1] command runs a number of self-consistency checks
1380 on the repository, and reports on any problems. This may take some
1381 time. The most common warning by far is about "dangling" objects:
1383 -------------------------------------------------
1385 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1386 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1387 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1388 dangling blob 218761f9d90712d37a9c5e36f406f92202db07eb
1389 dangling commit bf093535a34a4d35731aa2bd90fe6b176302f14f
1390 dangling commit 8e4bec7f2ddaa268bef999853c25755452100f8e
1391 dangling tree d50bb86186bf27b681d25af89d3b5b68382e4085
1392 dangling tree b24c2473f1fd3d91352a624795be026d64c8841f
1394 -------------------------------------------------
1396 Dangling objects are not a problem. At worst they may take up a little
1397 extra disk space. They can sometimes provide a last-resort method of
1398 recovery lost work--see <<dangling-objects>> for details. However, if
1399 you want, you may remove them with gitlink:git-prune[1] or the --prune
1400 option to gitlink:git-gc[1]:
1402 -------------------------------------------------
1404 -------------------------------------------------
1406 This may be time-consuming. Unlike most other git operations (including
1407 git-gc when run without any options), it is not safe to prune while
1408 other git operations are in progress in the same repository.
1410 [[recovering-lost-changes]]
1411 Recovering lost changes
1412 ~~~~~~~~~~~~~~~~~~~~~~~
1418 Say you modify a branch with gitlink:git-reset[1] --hard, and then
1419 realize that the branch was the only reference you had to that point in
1422 Fortunately, git also keeps a log, called a "reflog", of all the
1423 previous values of each branch. So in this case you can still find the
1424 old history using, for example,
1426 -------------------------------------------------
1427 $ git log master@{1}
1428 -------------------------------------------------
1430 This lists the commits reachable from the previous version of the head.
1431 This syntax can be used to with any git command that accepts a commit,
1432 not just with git log. Some other examples:
1434 -------------------------------------------------
1435 $ git show master@{2} # See where the branch pointed 2,
1436 $ git show master@{3} # 3, ... changes ago.
1437 $ gitk master@{yesterday} # See where it pointed yesterday,
1438 $ gitk master@{"1 week ago"} # ... or last week
1439 $ git log --walk-reflogs master # show reflog entries for master
1440 -------------------------------------------------
1442 A separate reflog is kept for the HEAD, so
1444 -------------------------------------------------
1445 $ git show HEAD@{"1 week ago"}
1446 -------------------------------------------------
1448 will show what HEAD pointed to one week ago, not what the current branch
1449 pointed to one week ago. This allows you to see the history of what
1452 The reflogs are kept by default for 30 days, after which they may be
1453 pruned. See gitlink:git-reflog[1] and gitlink:git-gc[1] to learn
1454 how to control this pruning, and see the "SPECIFYING REVISIONS"
1455 section of gitlink:git-rev-parse[1] for details.
1457 Note that the reflog history is very different from normal git history.
1458 While normal history is shared by every repository that works on the
1459 same project, the reflog history is not shared: it tells you only about
1460 how the branches in your local repository have changed over time.
1462 [[dangling-object-recovery]]
1463 Examining dangling objects
1464 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1466 In some situations the reflog may not be able to save you. For example,
1467 suppose you delete a branch, then realize you need the history it
1468 contained. The reflog is also deleted; however, if you have not yet
1469 pruned the repository, then you may still be able to find the lost
1470 commits in the dangling objects that git-fsck reports. See
1471 <<dangling-objects>> for the details.
1473 -------------------------------------------------
1475 dangling commit 7281251ddd2a61e38657c827739c57015671a6b3
1476 dangling commit 2706a059f258c6b245f298dc4ff2ccd30ec21a63
1477 dangling commit 13472b7c4b80851a1bc551779171dcb03655e9b5
1479 -------------------------------------------------
1482 one of those dangling commits with, for example,
1484 ------------------------------------------------
1485 $ gitk 7281251ddd --not --all
1486 ------------------------------------------------
1488 which does what it sounds like: it says that you want to see the commit
1489 history that is described by the dangling commit(s), but not the
1490 history that is described by all your existing branches and tags. Thus
1491 you get exactly the history reachable from that commit that is lost.
1492 (And notice that it might not be just one commit: we only report the
1493 "tip of the line" as being dangling, but there might be a whole deep
1494 and complex commit history that was dropped.)
1496 If you decide you want the history back, you can always create a new
1497 reference pointing to it, for example, a new branch:
1499 ------------------------------------------------
1500 $ git branch recovered-branch 7281251ddd
1501 ------------------------------------------------
1503 Other types of dangling objects (blobs and trees) are also possible, and
1504 dangling objects can arise in other situations.
1507 [[sharing-development]]
1508 Sharing development with others
1509 ===============================
1511 [[getting-updates-with-git-pull]]
1512 Getting updates with git pull
1513 -----------------------------
1515 After you clone a repository and make a few changes of your own, you
1516 may wish to check the original repository for updates and merge them
1519 We have already seen <<Updating-a-repository-with-git-fetch,how to
1520 keep remote tracking branches up to date>> with gitlink:git-fetch[1],
1521 and how to merge two branches. So you can merge in changes from the
1522 original repository's master branch with:
1524 -------------------------------------------------
1526 $ git merge origin/master
1527 -------------------------------------------------
1529 However, the gitlink:git-pull[1] command provides a way to do this in
1532 -------------------------------------------------
1533 $ git pull origin master
1534 -------------------------------------------------
1536 In fact, "origin" is normally the default repository to pull from,
1537 and the default branch is normally the HEAD of the remote repository,
1538 so often you can accomplish the above with just
1540 -------------------------------------------------
1542 -------------------------------------------------
1544 See the descriptions of the branch.<name>.remote and branch.<name>.merge
1545 options in gitlink:git-config[1] to learn how to control these defaults
1546 depending on the current branch. Also note that the --track option to
1547 gitlink:git-branch[1] and gitlink:git-checkout[1] can be used to
1548 automatically set the default remote branch to pull from at the time
1549 that a branch is created:
1551 -------------------------------------------------
1552 $ git checkout --track -b origin/maint maint
1553 -------------------------------------------------
1555 In addition to saving you keystrokes, "git pull" also helps you by
1556 producing a default commit message documenting the branch and
1557 repository that you pulled from.
1559 (But note that no such commit will be created in the case of a
1560 <<fast-forwards,fast forward>>; instead, your branch will just be
1561 updated to point to the latest commit from the upstream branch.)
1563 The git-pull command can also be given "." as the "remote" repository,
1564 in which case it just merges in a branch from the current repository; so
1567 -------------------------------------------------
1570 -------------------------------------------------
1572 are roughly equivalent. The former is actually very commonly used.
1574 [[submitting-patches]]
1575 Submitting patches to a project
1576 -------------------------------
1578 If you just have a few changes, the simplest way to submit them may
1579 just be to send them as patches in email:
1581 First, use gitlink:git-format-patch[1]; for example:
1583 -------------------------------------------------
1584 $ git format-patch origin
1585 -------------------------------------------------
1587 will produce a numbered series of files in the current directory, one
1588 for each patch in the current branch but not in origin/HEAD.
1590 You can then import these into your mail client and send them by
1591 hand. However, if you have a lot to send at once, you may prefer to
1592 use the gitlink:git-send-email[1] script to automate the process.
1593 Consult the mailing list for your project first to determine how they
1594 prefer such patches be handled.
1596 [[importing-patches]]
1597 Importing patches to a project
1598 ------------------------------
1600 Git also provides a tool called gitlink:git-am[1] (am stands for
1601 "apply mailbox"), for importing such an emailed series of patches.
1602 Just save all of the patch-containing messages, in order, into a
1603 single mailbox file, say "patches.mbox", then run
1605 -------------------------------------------------
1606 $ git am -3 patches.mbox
1607 -------------------------------------------------
1609 Git will apply each patch in order; if any conflicts are found, it
1610 will stop, and you can fix the conflicts as described in
1611 "<<resolving-a-merge,Resolving a merge>>". (The "-3" option tells
1612 git to perform a merge; if you would prefer it just to abort and
1613 leave your tree and index untouched, you may omit that option.)
1615 Once the index is updated with the results of the conflict
1616 resolution, instead of creating a new commit, just run
1618 -------------------------------------------------
1620 -------------------------------------------------
1622 and git will create the commit for you and continue applying the
1623 remaining patches from the mailbox.
1625 The final result will be a series of commits, one for each patch in
1626 the original mailbox, with authorship and commit log message each
1627 taken from the message containing each patch.
1629 [[setting-up-a-public-repository]]
1630 Setting up a public repository
1631 ------------------------------
1633 Another way to submit changes to a project is to simply tell the
1634 maintainer of that project to pull from your repository, exactly as
1635 you did in the section "<<getting-updates-with-git-pull, Getting
1636 updates with git pull>>".
1638 If you and maintainer both have accounts on the same machine, then
1639 then you can just pull changes from each other's repositories
1640 directly; note that all of the commands (gitlink:git-clone[1],
1641 git-fetch[1], git-pull[1], etc.) that accept a URL as an argument
1642 will also accept a local directory name; so, for example, you can
1645 -------------------------------------------------
1646 $ git clone /path/to/repository
1647 $ git pull /path/to/other/repository
1648 -------------------------------------------------
1650 If this sort of setup is inconvenient or impossible, another (more
1651 common) option is to set up a public repository on a public server.
1652 This also allows you to cleanly separate private work in progress
1653 from publicly visible work.
1655 You will continue to do your day-to-day work in your personal
1656 repository, but periodically "push" changes from your personal
1657 repository into your public repository, allowing other developers to
1658 pull from that repository. So the flow of changes, in a situation
1659 where there is one other developer with a public repository, looks
1663 your personal repo ------------------> your public repo
1666 | you pull | they pull
1670 their public repo <------------------- their repo
1672 Now, assume your personal repository is in the directory ~/proj. We
1673 first create a new clone of the repository:
1675 -------------------------------------------------
1676 $ git clone --bare ~/proj proj.git
1677 -------------------------------------------------
1679 The resulting directory proj.git contains a "bare" git repository--it is
1680 just the contents of the ".git" directory, without a checked-out copy of
1681 a working directory.
1683 Next, copy proj.git to the server where you plan to host the
1684 public repository. You can use scp, rsync, or whatever is most
1687 If somebody else maintains the public server, they may already have
1688 set up a git service for you, and you may skip to the section
1689 "<<pushing-changes-to-a-public-repository,Pushing changes to a public
1690 repository>>", below.
1692 Otherwise, the following sections explain how to export your newly
1693 created public repository:
1695 [[exporting-via-http]]
1696 Exporting a git repository via http
1697 -----------------------------------
1699 The git protocol gives better performance and reliability, but on a
1700 host with a web server set up, http exports may be simpler to set up.
1702 All you need to do is place the newly created bare git repository in
1703 a directory that is exported by the web server, and make some
1704 adjustments to give web clients some extra information they need:
1706 -------------------------------------------------
1707 $ mv proj.git /home/you/public_html/proj.git
1709 $ git --bare update-server-info
1710 $ chmod a+x hooks/post-update
1711 -------------------------------------------------
1713 (For an explanation of the last two lines, see
1714 gitlink:git-update-server-info[1], and the documentation
1715 link:hooks.txt[Hooks used by git].)
1717 Advertise the url of proj.git. Anybody else should then be able to
1718 clone or pull from that url, for example with a commandline like:
1720 -------------------------------------------------
1721 $ git clone http://yourserver.com/~you/proj.git
1722 -------------------------------------------------
1725 link:howto/setup-git-server-over-http.txt[setup-git-server-over-http]
1726 for a slightly more sophisticated setup using WebDAV which also
1727 allows pushing over http.)
1729 [[exporting-via-git]]
1730 Exporting a git repository via the git protocol
1731 -----------------------------------------------
1733 This is the preferred method.
1735 For now, we refer you to the gitlink:git-daemon[1] man page for
1736 instructions. (See especially the examples section.)
1738 [[pushing-changes-to-a-public-repository]]
1739 Pushing changes to a public repository
1740 --------------------------------------
1742 Note that the two techniques outline above (exporting via
1743 <<exporting-via-http,http>> or <<exporting-via-git,git>>) allow other
1744 maintainers to fetch your latest changes, but they do not allow write
1745 access, which you will need to update the public repository with the
1746 latest changes created in your private repository.
1748 The simplest way to do this is using gitlink:git-push[1] and ssh; to
1749 update the remote branch named "master" with the latest state of your
1750 branch named "master", run
1752 -------------------------------------------------
1753 $ git push ssh://yourserver.com/~you/proj.git master:master
1754 -------------------------------------------------
1758 -------------------------------------------------
1759 $ git push ssh://yourserver.com/~you/proj.git master
1760 -------------------------------------------------
1762 As with git-fetch, git-push will complain if this does not result in
1763 a <<fast-forwards,fast forward>>. Normally this is a sign of
1764 something wrong. However, if you are sure you know what you're
1765 doing, you may force git-push to perform the update anyway by
1766 proceeding the branch name by a plus sign:
1768 -------------------------------------------------
1769 $ git push ssh://yourserver.com/~you/proj.git +master
1770 -------------------------------------------------
1772 As with git-fetch, you may also set up configuration options to
1773 save typing; so, for example, after
1775 -------------------------------------------------
1776 $ cat >>.git/config <<EOF
1777 [remote "public-repo"]
1778 url = ssh://yourserver.com/~you/proj.git
1780 -------------------------------------------------
1782 you should be able to perform the above push with just
1784 -------------------------------------------------
1785 $ git push public-repo master
1786 -------------------------------------------------
1788 See the explanations of the remote.<name>.url, branch.<name>.remote,
1789 and remote.<name>.push options in gitlink:git-config[1] for
1792 [[setting-up-a-shared-repository]]
1793 Setting up a shared repository
1794 ------------------------------
1796 Another way to collaborate is by using a model similar to that
1797 commonly used in CVS, where several developers with special rights
1798 all push to and pull from a single shared repository. See
1799 link:cvs-migration.txt[git for CVS users] for instructions on how to
1802 [[setting-up-gitweb]]
1803 Allow web browsing of a repository
1804 ----------------------------------
1806 The gitweb cgi script provides users an easy way to browse your
1807 project's files and history without having to install git; see the file
1808 gitweb/INSTALL in the git source tree for instructions on setting it up.
1810 [[sharing-development-examples]]
1814 [[maintaining-topic-branches]]
1815 Maintaining topic branches for a Linux subsystem maintainer
1816 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1818 This describes how Tony Luck uses git in his role as maintainer of the
1819 IA64 architecture for the Linux kernel.
1821 He uses two public branches:
1823 - A "test" tree into which patches are initially placed so that they
1824 can get some exposure when integrated with other ongoing development.
1825 This tree is available to Andrew for pulling into -mm whenever he
1828 - A "release" tree into which tested patches are moved for final sanity
1829 checking, and as a vehicle to send them upstream to Linus (by sending
1830 him a "please pull" request.)
1832 He also uses a set of temporary branches ("topic branches"), each
1833 containing a logical grouping of patches.
1835 To set this up, first create your work tree by cloning Linus's public
1838 -------------------------------------------------
1839 $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git work
1841 -------------------------------------------------
1843 Linus's tree will be stored in the remote branch named origin/master,
1844 and can be updated using gitlink:git-fetch[1]; you can track other
1845 public trees using gitlink:git-remote[1] to set up a "remote" and
1846 git-fetch[1] to keep them up-to-date; see <<repositories-and-branches>>.
1848 Now create the branches in which you are going to work; these start out
1849 at the current tip of origin/master branch, and should be set up (using
1850 the --track option to gitlink:git-branch[1]) to merge changes in from
1853 -------------------------------------------------
1854 $ git branch --track test origin/master
1855 $ git branch --track release origin/master
1856 -------------------------------------------------
1858 These can be easily kept up to date using gitlink:git-pull[1]
1860 -------------------------------------------------
1861 $ git checkout test && git pull
1862 $ git checkout release && git pull
1863 -------------------------------------------------
1865 Important note! If you have any local changes in these branches, then
1866 this merge will create a commit object in the history (with no local
1867 changes git will simply do a "Fast forward" merge). Many people dislike
1868 the "noise" that this creates in the Linux history, so you should avoid
1869 doing this capriciously in the "release" branch, as these noisy commits
1870 will become part of the permanent history when you ask Linus to pull
1871 from the release branch.
1873 A few configuration variables (see gitlink:git-config[1]) can
1874 make it easy to push both branches to your public tree. (See
1875 <<setting-up-a-public-repository>>.)
1877 -------------------------------------------------
1878 $ cat >> .git/config <<EOF
1880 url = master.kernel.org:/pub/scm/linux/kernel/git/aegl/linux-2.6.git
1884 -------------------------------------------------
1886 Then you can push both the test and release trees using
1887 gitlink:git-push[1]:
1889 -------------------------------------------------
1891 -------------------------------------------------
1893 or push just one of the test and release branches using:
1895 -------------------------------------------------
1896 $ git push mytree test
1897 -------------------------------------------------
1901 -------------------------------------------------
1902 $ git push mytree release
1903 -------------------------------------------------
1905 Now to apply some patches from the community. Think of a short
1906 snappy name for a branch to hold this patch (or related group of
1907 patches), and create a new branch from the current tip of Linus's
1910 -------------------------------------------------
1911 $ git checkout -b speed-up-spinlocks origin
1912 -------------------------------------------------
1914 Now you apply the patch(es), run some tests, and commit the change(s). If
1915 the patch is a multi-part series, then you should apply each as a separate
1916 commit to this branch.
1918 -------------------------------------------------
1919 $ ... patch ... test ... commit [ ... patch ... test ... commit ]*
1920 -------------------------------------------------
1922 When you are happy with the state of this change, you can pull it into the
1923 "test" branch in preparation to make it public:
1925 -------------------------------------------------
1926 $ git checkout test && git pull . speed-up-spinlocks
1927 -------------------------------------------------
1929 It is unlikely that you would have any conflicts here ... but you might if you
1930 spent a while on this step and had also pulled new versions from upstream.
1932 Some time later when enough time has passed and testing done, you can pull the
1933 same branch into the "release" tree ready to go upstream. This is where you
1934 see the value of keeping each patch (or patch series) in its own branch. It
1935 means that the patches can be moved into the "release" tree in any order.
1937 -------------------------------------------------
1938 $ git checkout release && git pull . speed-up-spinlocks
1939 -------------------------------------------------
1941 After a while, you will have a number of branches, and despite the
1942 well chosen names you picked for each of them, you may forget what
1943 they are for, or what status they are in. To get a reminder of what
1944 changes are in a specific branch, use:
1946 -------------------------------------------------
1947 $ git log linux..branchname | git-shortlog
1948 -------------------------------------------------
1950 To see whether it has already been merged into the test or release branches
1953 -------------------------------------------------
1954 $ git log test..branchname
1955 -------------------------------------------------
1959 -------------------------------------------------
1960 $ git log release..branchname
1961 -------------------------------------------------
1963 (If this branch has not yet been merged you will see some log entries.
1964 If it has been merged, then there will be no output.)
1966 Once a patch completes the great cycle (moving from test to release,
1967 then pulled by Linus, and finally coming back into your local
1968 "origin/master" branch) the branch for this change is no longer needed.
1969 You detect this when the output from:
1971 -------------------------------------------------
1972 $ git log origin..branchname
1973 -------------------------------------------------
1975 is empty. At this point the branch can be deleted:
1977 -------------------------------------------------
1978 $ git branch -d branchname
1979 -------------------------------------------------
1981 Some changes are so trivial that it is not necessary to create a separate
1982 branch and then merge into each of the test and release branches. For
1983 these changes, just apply directly to the "release" branch, and then
1984 merge that into the "test" branch.
1986 To create diffstat and shortlog summaries of changes to include in a "please
1987 pull" request to Linus you can use:
1989 -------------------------------------------------
1990 $ git diff --stat origin..release
1991 -------------------------------------------------
1995 -------------------------------------------------
1996 $ git log -p origin..release | git shortlog
1997 -------------------------------------------------
1999 Here are some of the scripts that simplify all this even further.
2001 -------------------------------------------------
2002 ==== update script ====
2003 # Update a branch in my GIT tree. If the branch to be updated
2004 # is origin, then pull from kernel.org. Otherwise merge
2005 # origin/master branch into test|release branch
2009 git checkout $1 && git pull . origin
2012 before=$(cat .git/refs/remotes/origin/master)
2014 after=$(cat .git/refs/remotes/origin/master)
2015 if [ $before != $after ]
2017 git log $before..$after | git shortlog
2021 echo "Usage: $0 origin|test|release" 1>&2
2025 -------------------------------------------------
2027 -------------------------------------------------
2028 ==== merge script ====
2029 # Merge a branch into either the test or release branch
2035 echo "Usage: $pname branch test|release" 1>&2
2039 if [ ! -f .git/refs/heads/"$1" ]
2041 echo "Can't see branch <$1>" 1>&2
2047 if [ $(git log $2..$1 | wc -c) -eq 0 ]
2049 echo $1 already merged into $2 1>&2
2052 git checkout $2 && git pull . $1
2058 -------------------------------------------------
2060 -------------------------------------------------
2061 ==== status script ====
2062 # report on status of my ia64 GIT tree
2066 restore=$(tput setab 9)
2068 if [ `git rev-list test..release | wc -c` -gt 0 ]
2070 echo $rb Warning: commits in release that are not in test $restore
2071 git log test..release
2074 for branch in `ls .git/refs/heads`
2076 if [ $branch = test -o $branch = release ]
2081 echo -n $gb ======= $branch ====== $restore " "
2083 for ref in test release origin/master
2085 if [ `git rev-list $ref..$branch | wc -c` -gt 0 ]
2087 status=$status${ref:0:1}
2092 echo $rb Need to pull into test $restore
2098 echo "Waiting for linus"
2101 echo $rb All done $restore
2104 echo $rb "<$status>" $restore
2107 git log origin/master..$branch | git shortlog
2109 -------------------------------------------------
2112 [[cleaning-up-history]]
2113 Rewriting history and maintaining patch series
2114 ==============================================
2116 Normally commits are only added to a project, never taken away or
2117 replaced. Git is designed with this assumption, and violating it will
2118 cause git's merge machinery (for example) to do the wrong thing.
2120 However, there is a situation in which it can be useful to violate this
2124 Creating the perfect patch series
2125 ---------------------------------
2127 Suppose you are a contributor to a large project, and you want to add a
2128 complicated feature, and to present it to the other developers in a way
2129 that makes it easy for them to read your changes, verify that they are
2130 correct, and understand why you made each change.
2132 If you present all of your changes as a single patch (or commit), they
2133 may find that it is too much to digest all at once.
2135 If you present them with the entire history of your work, complete with
2136 mistakes, corrections, and dead ends, they may be overwhelmed.
2138 So the ideal is usually to produce a series of patches such that:
2140 1. Each patch can be applied in order.
2142 2. Each patch includes a single logical change, together with a
2143 message explaining the change.
2145 3. No patch introduces a regression: after applying any initial
2146 part of the series, the resulting project still compiles and
2147 works, and has no bugs that it didn't have before.
2149 4. The complete series produces the same end result as your own
2150 (probably much messier!) development process did.
2152 We will introduce some tools that can help you do this, explain how to
2153 use them, and then explain some of the problems that can arise because
2154 you are rewriting history.
2156 [[using-git-rebase]]
2157 Keeping a patch series up to date using git-rebase
2158 --------------------------------------------------
2160 Suppose that you create a branch "mywork" on a remote-tracking branch
2161 "origin", and create some commits on top of it:
2163 -------------------------------------------------
2164 $ git checkout -b mywork origin
2170 -------------------------------------------------
2172 You have performed no merges into mywork, so it is just a simple linear
2173 sequence of patches on top of "origin":
2175 ................................................
2179 ................................................
2181 Some more interesting work has been done in the upstream project, and
2182 "origin" has advanced:
2184 ................................................
2185 o--o--O--o--o--o <-- origin
2188 ................................................
2190 At this point, you could use "pull" to merge your changes back in;
2191 the result would create a new merge commit, like this:
2193 ................................................
2194 o--o--O--o--o--o <-- origin
2196 a--b--c--m <-- mywork
2197 ................................................
2199 However, if you prefer to keep the history in mywork a simple series of
2200 commits without any merges, you may instead choose to use
2201 gitlink:git-rebase[1]:
2203 -------------------------------------------------
2204 $ git checkout mywork
2206 -------------------------------------------------
2208 This will remove each of your commits from mywork, temporarily saving
2209 them as patches (in a directory named ".dotest"), update mywork to
2210 point at the latest version of origin, then apply each of the saved
2211 patches to the new mywork. The result will look like:
2214 ................................................
2215 o--o--O--o--o--o <-- origin
2217 a'--b'--c' <-- mywork
2218 ................................................
2220 In the process, it may discover conflicts. In that case it will stop
2221 and allow you to fix the conflicts; after fixing conflicts, use "git
2222 add" to update the index with those contents, and then, instead of
2223 running git-commit, just run
2225 -------------------------------------------------
2226 $ git rebase --continue
2227 -------------------------------------------------
2229 and git will continue applying the rest of the patches.
2231 At any point you may use the --abort option to abort this process and
2232 return mywork to the state it had before you started the rebase:
2234 -------------------------------------------------
2235 $ git rebase --abort
2236 -------------------------------------------------
2238 [[modifying-one-commit]]
2239 Modifying a single commit
2240 -------------------------
2242 We saw in <<fixing-a-mistake-by-editing-history>> that you can replace the
2243 most recent commit using
2245 -------------------------------------------------
2246 $ git commit --amend
2247 -------------------------------------------------
2249 which will replace the old commit by a new commit incorporating your
2250 changes, giving you a chance to edit the old commit message first.
2252 You can also use a combination of this and gitlink:git-rebase[1] to edit
2253 commits further back in your history. First, tag the problematic commit with
2255 -------------------------------------------------
2256 $ git tag bad mywork~5
2257 -------------------------------------------------
2259 (Either gitk or git-log may be useful for finding the commit.)
2261 Then check out that commit, edit it, and rebase the rest of the series
2262 on top of it (note that we could check out the commit on a temporary
2263 branch, but instead we're using a <<detached-head,detached head>>):
2265 -------------------------------------------------
2267 $ # make changes here and update the index
2268 $ git commit --amend
2269 $ git rebase --onto HEAD bad mywork
2270 -------------------------------------------------
2272 When you're done, you'll be left with mywork checked out, with the top
2273 patches on mywork reapplied on top of your modified commit. You can
2276 -------------------------------------------------
2278 -------------------------------------------------
2280 Note that the immutable nature of git history means that you haven't really
2281 "modified" existing commits; instead, you have replaced the old commits with
2282 new commits having new object names.
2284 [[reordering-patch-series]]
2285 Reordering or selecting from a patch series
2286 -------------------------------------------
2288 Given one existing commit, the gitlink:git-cherry-pick[1] command
2289 allows you to apply the change introduced by that commit and create a
2290 new commit that records it. So, for example, if "mywork" points to a
2291 series of patches on top of "origin", you might do something like:
2293 -------------------------------------------------
2294 $ git checkout -b mywork-new origin
2295 $ gitk origin..mywork &
2296 -------------------------------------------------
2298 And browse through the list of patches in the mywork branch using gitk,
2299 applying them (possibly in a different order) to mywork-new using
2300 cherry-pick, and possibly modifying them as you go using commit
2303 Another technique is to use git-format-patch to create a series of
2304 patches, then reset the state to before the patches:
2306 -------------------------------------------------
2307 $ git format-patch origin
2308 $ git reset --hard origin
2309 -------------------------------------------------
2311 Then modify, reorder, or eliminate patches as preferred before applying
2312 them again with gitlink:git-am[1].
2314 [[patch-series-tools]]
2318 There are numerous other tools, such as stgit, which exist for the
2319 purpose of maintaining a patch series. These are outside of the scope of
2322 [[problems-with-rewriting-history]]
2323 Problems with rewriting history
2324 -------------------------------
2326 The primary problem with rewriting the history of a branch has to do
2327 with merging. Suppose somebody fetches your branch and merges it into
2328 their branch, with a result something like this:
2330 ................................................
2331 o--o--O--o--o--o <-- origin
2333 t--t--t--m <-- their branch:
2334 ................................................
2336 Then suppose you modify the last three commits:
2338 ................................................
2339 o--o--o <-- new head of origin
2341 o--o--O--o--o--o <-- old head of origin
2342 ................................................
2344 If we examined all this history together in one repository, it will
2347 ................................................
2348 o--o--o <-- new head of origin
2350 o--o--O--o--o--o <-- old head of origin
2352 t--t--t--m <-- their branch:
2353 ................................................
2355 Git has no way of knowing that the new head is an updated version of
2356 the old head; it treats this situation exactly the same as it would if
2357 two developers had independently done the work on the old and new heads
2358 in parallel. At this point, if someone attempts to merge the new head
2359 in to their branch, git will attempt to merge together the two (old and
2360 new) lines of development, instead of trying to replace the old by the
2361 new. The results are likely to be unexpected.
2363 You may still choose to publish branches whose history is rewritten,
2364 and it may be useful for others to be able to fetch those branches in
2365 order to examine or test them, but they should not attempt to pull such
2366 branches into their own work.
2368 For true distributed development that supports proper merging,
2369 published branches should never be rewritten.
2371 [[advanced-branch-management]]
2372 Advanced branch management
2373 ==========================
2375 [[fetching-individual-branches]]
2376 Fetching individual branches
2377 ----------------------------
2379 Instead of using gitlink:git-remote[1], you can also choose just
2380 to update one branch at a time, and to store it locally under an
2383 -------------------------------------------------
2384 $ git fetch origin todo:my-todo-work
2385 -------------------------------------------------
2387 The first argument, "origin", just tells git to fetch from the
2388 repository you originally cloned from. The second argument tells git
2389 to fetch the branch named "todo" from the remote repository, and to
2390 store it locally under the name refs/heads/my-todo-work.
2392 You can also fetch branches from other repositories; so
2394 -------------------------------------------------
2395 $ git fetch git://example.com/proj.git master:example-master
2396 -------------------------------------------------
2398 will create a new branch named "example-master" and store in it the
2399 branch named "master" from the repository at the given URL. If you
2400 already have a branch named example-master, it will attempt to
2401 <<fast-forwards,fast-forward>> to the commit given by example.com's
2402 master branch. In more detail:
2404 [[fetch-fast-forwards]]
2405 git fetch and fast-forwards
2406 ---------------------------
2408 In the previous example, when updating an existing branch, "git
2409 fetch" checks to make sure that the most recent commit on the remote
2410 branch is a descendant of the most recent commit on your copy of the
2411 branch before updating your copy of the branch to point at the new
2412 commit. Git calls this process a <<fast-forwards,fast forward>>.
2414 A fast forward looks something like this:
2416 ................................................
2417 o--o--o--o <-- old head of the branch
2419 o--o--o <-- new head of the branch
2420 ................................................
2423 In some cases it is possible that the new head will *not* actually be
2424 a descendant of the old head. For example, the developer may have
2425 realized she made a serious mistake, and decided to backtrack,
2426 resulting in a situation like:
2428 ................................................
2429 o--o--o--o--a--b <-- old head of the branch
2431 o--o--o <-- new head of the branch
2432 ................................................
2434 In this case, "git fetch" will fail, and print out a warning.
2436 In that case, you can still force git to update to the new head, as
2437 described in the following section. However, note that in the
2438 situation above this may mean losing the commits labeled "a" and "b",
2439 unless you've already created a reference of your own pointing to
2443 Forcing git fetch to do non-fast-forward updates
2444 ------------------------------------------------
2446 If git fetch fails because the new head of a branch is not a
2447 descendant of the old head, you may force the update with:
2449 -------------------------------------------------
2450 $ git fetch git://example.com/proj.git +master:refs/remotes/example/master
2451 -------------------------------------------------
2453 Note the addition of the "+" sign. Alternatively, you can use the "-f"
2454 flag to force updates of all the fetched branches, as in:
2456 -------------------------------------------------
2457 $ git fetch -f origin
2458 -------------------------------------------------
2460 Be aware that commits that the old version of example/master pointed at
2461 may be lost, as we saw in the previous section.
2463 [[remote-branch-configuration]]
2464 Configuring remote branches
2465 ---------------------------
2467 We saw above that "origin" is just a shortcut to refer to the
2468 repository that you originally cloned from. This information is
2469 stored in git configuration variables, which you can see using
2470 gitlink:git-config[1]:
2472 -------------------------------------------------
2474 core.repositoryformatversion=0
2476 core.logallrefupdates=true
2477 remote.origin.url=git://git.kernel.org/pub/scm/git/git.git
2478 remote.origin.fetch=+refs/heads/*:refs/remotes/origin/*
2479 branch.master.remote=origin
2480 branch.master.merge=refs/heads/master
2481 -------------------------------------------------
2483 If there are other repositories that you also use frequently, you can
2484 create similar configuration options to save typing; for example,
2487 -------------------------------------------------
2488 $ git config remote.example.url git://example.com/proj.git
2489 -------------------------------------------------
2491 then the following two commands will do the same thing:
2493 -------------------------------------------------
2494 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2495 $ git fetch example master:refs/remotes/example/master
2496 -------------------------------------------------
2498 Even better, if you add one more option:
2500 -------------------------------------------------
2501 $ git config remote.example.fetch master:refs/remotes/example/master
2502 -------------------------------------------------
2504 then the following commands will all do the same thing:
2506 -------------------------------------------------
2507 $ git fetch git://example.com/proj.git master:refs/remotes/example/master
2508 $ git fetch example master:refs/remotes/example/master
2510 -------------------------------------------------
2512 You can also add a "+" to force the update each time:
2514 -------------------------------------------------
2515 $ git config remote.example.fetch +master:ref/remotes/example/master
2516 -------------------------------------------------
2518 Don't do this unless you're sure you won't mind "git fetch" possibly
2519 throwing away commits on mybranch.
2521 Also note that all of the above configuration can be performed by
2522 directly editing the file .git/config instead of using
2523 gitlink:git-config[1].
2525 See gitlink:git-config[1] for more details on the configuration
2526 options mentioned above.
2533 Git depends on two fundamental abstractions: the "object database", and
2534 the "current directory cache" aka "index".
2536 [[the-object-database]]
2540 The object database is literally just a content-addressable collection
2541 of objects. All objects are named by their content, which is
2542 approximated by the SHA1 hash of the object itself. Objects may refer
2543 to other objects (by referencing their SHA1 hash), and so you can
2544 build up a hierarchy of objects.
2546 All objects have a statically determined "type" which is
2547 determined at object creation time, and which identifies the format of
2548 the object (i.e. how it is used, and how it can refer to other
2549 objects). There are currently four different object types: "blob",
2550 "tree", "commit", and "tag".
2552 A <<def_blob_object,"blob" object>> cannot refer to any other object,
2553 and is, as the name implies, a pure storage object containing some
2554 user data. It is used to actually store the file data, i.e. a blob
2555 object is associated with some particular version of some file.
2557 A <<def_tree_object,"tree" object>> is an object that ties one or more
2558 "blob" objects into a directory structure. In addition, a tree object
2559 can refer to other tree objects, thus creating a directory hierarchy.
2561 A <<def_commit_object,"commit" object>> ties such directory hierarchies
2562 together into a <<def_DAG,directed acyclic graph>> of revisions - each
2563 "commit" is associated with exactly one tree (the directory hierarchy at
2564 the time of the commit). In addition, a "commit" refers to one or more
2565 "parent" commit objects that describe the history of how we arrived at
2566 that directory hierarchy.
2568 As a special case, a commit object with no parents is called the "root"
2569 commit, and is the point of an initial project commit. Each project
2570 must have at least one root, and while you can tie several different
2571 root objects together into one project by creating a commit object which
2572 has two or more separate roots as its ultimate parents, that's probably
2573 just going to confuse people. So aim for the notion of "one root object
2574 per project", even if git itself does not enforce that.
2576 A <<def_tag_object,"tag" object>> symbolically identifies and can be
2577 used to sign other objects. It contains the identifier and type of
2578 another object, a symbolic name (of course!) and, optionally, a
2581 Regardless of object type, all objects share the following
2582 characteristics: they are all deflated with zlib, and have a header
2583 that not only specifies their type, but also provides size information
2584 about the data in the object. It's worth noting that the SHA1 hash
2585 that is used to name the object is the hash of the original data
2586 plus this header, so `sha1sum` 'file' does not match the object name
2588 (Historical note: in the dawn of the age of git the hash
2589 was the sha1 of the 'compressed' object.)
2591 As a result, the general consistency of an object can always be tested
2592 independently of the contents or the type of the object: all objects can
2593 be validated by verifying that (a) their hashes match the content of the
2594 file and (b) the object successfully inflates to a stream of bytes that
2595 forms a sequence of <ascii type without space> + <space> + <ascii decimal
2596 size> + <byte\0> + <binary object data>.
2598 The structured objects can further have their structure and
2599 connectivity to other objects verified. This is generally done with
2600 the `git-fsck` program, which generates a full dependency graph
2601 of all objects, and verifies their internal consistency (in addition
2602 to just verifying their superficial consistency through the hash).
2604 The object types in some more detail:
2610 A "blob" object is nothing but a binary blob of data, and doesn't
2611 refer to anything else. There is no signature or any other
2612 verification of the data, so while the object is consistent (it 'is'
2613 indexed by its sha1 hash, so the data itself is certainly correct), it
2614 has absolutely no other attributes. No name associations, no
2615 permissions. It is purely a blob of data (i.e. normally "file
2618 In particular, since the blob is entirely defined by its data, if two
2619 files in a directory tree (or in multiple different versions of the
2620 repository) have the same contents, they will share the same blob
2621 object. The object is totally independent of its location in the
2622 directory tree, and renaming a file does not change the object that
2623 file is associated with in any way.
2625 A blob is typically created when gitlink:git-update-index[1]
2626 is run, and its data can be accessed by gitlink:git-cat-file[1].
2632 The next hierarchical object type is the "tree" object. A tree object
2633 is a list of mode/name/blob data, sorted by name. Alternatively, the
2634 mode data may specify a directory mode, in which case instead of
2635 naming a blob, that name is associated with another TREE object.
2637 Like the "blob" object, a tree object is uniquely determined by the
2638 set contents, and so two separate but identical trees will always
2639 share the exact same object. This is true at all levels, i.e. it's
2640 true for a "leaf" tree (which does not refer to any other trees, only
2641 blobs) as well as for a whole subdirectory.
2643 For that reason a "tree" object is just a pure data abstraction: it
2644 has no history, no signatures, no verification of validity, except
2645 that since the contents are again protected by the hash itself, we can
2646 trust that the tree is immutable and its contents never change.
2648 So you can trust the contents of a tree to be valid, the same way you
2649 can trust the contents of a blob, but you don't know where those
2650 contents 'came' from.
2652 Side note on trees: since a "tree" object is a sorted list of
2653 "filename+content", you can create a diff between two trees without
2654 actually having to unpack two trees. Just ignore all common parts,
2655 and your diff will look right. In other words, you can effectively
2656 (and efficiently) tell the difference between any two random trees by
2657 O(n) where "n" is the size of the difference, rather than the size of
2660 Side note 2 on trees: since the name of a "blob" depends entirely and
2661 exclusively on its contents (i.e. there are no names or permissions
2662 involved), you can see trivial renames or permission changes by
2663 noticing that the blob stayed the same. However, renames with data
2664 changes need a smarter "diff" implementation.
2666 A tree is created with gitlink:git-write-tree[1] and
2667 its data can be accessed by gitlink:git-ls-tree[1].
2668 Two trees can be compared with gitlink:git-diff-tree[1].
2674 The "commit" object is an object that introduces the notion of
2675 history into the picture. In contrast to the other objects, it
2676 doesn't just describe the physical state of a tree, it describes how
2677 we got there, and why.
2679 A "commit" is defined by the tree-object that it results in, the
2680 parent commits (zero, one or more) that led up to that point, and a
2681 comment on what happened. Again, a commit is not trusted per se:
2682 the contents are well-defined and "safe" due to the cryptographically
2683 strong signatures at all levels, but there is no reason to believe
2684 that the tree is "good" or that the merge information makes sense.
2685 The parents do not have to actually have any relationship with the
2686 result, for example.
2688 Note on commits: unlike some SCM's, commits do not contain
2689 rename information or file mode change information. All of that is
2690 implicit in the trees involved (the result tree, and the result trees
2691 of the parents), and describing that makes no sense in this idiotic
2694 A commit is created with gitlink:git-commit-tree[1] and
2695 its data can be accessed by gitlink:git-cat-file[1].
2701 An aside on the notion of "trust". Trust is really outside the scope
2702 of "git", but it's worth noting a few things. First off, since
2703 everything is hashed with SHA1, you 'can' trust that an object is
2704 intact and has not been messed with by external sources. So the name
2705 of an object uniquely identifies a known state - just not a state that
2706 you may want to trust.
2708 Furthermore, since the SHA1 signature of a commit refers to the
2709 SHA1 signatures of the tree it is associated with and the signatures
2710 of the parent, a single named commit specifies uniquely a whole set
2711 of history, with full contents. You can't later fake any step of the
2712 way once you have the name of a commit.
2714 So to introduce some real trust in the system, the only thing you need
2715 to do is to digitally sign just 'one' special note, which includes the
2716 name of a top-level commit. Your digital signature shows others
2717 that you trust that commit, and the immutability of the history of
2718 commits tells others that they can trust the whole history.
2720 In other words, you can easily validate a whole archive by just
2721 sending out a single email that tells the people the name (SHA1 hash)
2722 of the top commit, and digitally sign that email using something
2725 To assist in this, git also provides the tag object...
2731 Git provides the "tag" object to simplify creating, managing and
2732 exchanging symbolic and signed tokens. The "tag" object at its
2733 simplest simply symbolically identifies another object by containing
2734 the sha1, type and symbolic name.
2736 However it can optionally contain additional signature information
2737 (which git doesn't care about as long as there's less than 8k of
2738 it). This can then be verified externally to git.
2740 Note that despite the tag features, "git" itself only handles content
2741 integrity; the trust framework (and signature provision and
2742 verification) has to come from outside.
2744 A tag is created with gitlink:git-mktag[1],
2745 its data can be accessed by gitlink:git-cat-file[1],
2746 and the signature can be verified by
2747 gitlink:git-verify-tag[1].
2751 The "index" aka "Current Directory Cache"
2752 -----------------------------------------
2754 The index is a simple binary file, which contains an efficient
2755 representation of the contents of a virtual directory. It
2756 does so by a simple array that associates a set of names, dates,
2757 permissions and content (aka "blob") objects together. The cache is
2758 always kept ordered by name, and names are unique (with a few very
2759 specific rules) at any point in time, but the cache has no long-term
2760 meaning, and can be partially updated at any time.
2762 In particular, the index certainly does not need to be consistent with
2763 the current directory contents (in fact, most operations will depend on
2764 different ways to make the index 'not' be consistent with the directory
2765 hierarchy), but it has three very important attributes:
2767 '(a) it can re-generate the full state it caches (not just the
2768 directory structure: it contains pointers to the "blob" objects so
2769 that it can regenerate the data too)'
2771 As a special case, there is a clear and unambiguous one-way mapping
2772 from a current directory cache to a "tree object", which can be
2773 efficiently created from just the current directory cache without
2774 actually looking at any other data. So a directory cache at any one
2775 time uniquely specifies one and only one "tree" object (but has
2776 additional data to make it easy to match up that tree object with what
2777 has happened in the directory)
2779 '(b) it has efficient methods for finding inconsistencies between that
2780 cached state ("tree object waiting to be instantiated") and the
2783 '(c) it can additionally efficiently represent information about merge
2784 conflicts between different tree objects, allowing each pathname to be
2785 associated with sufficient information about the trees involved that
2786 you can create a three-way merge between them.'
2788 Those are the ONLY three things that the directory cache does. It's a
2789 cache, and the normal operation is to re-generate it completely from a
2790 known tree object, or update/compare it with a live tree that is being
2791 developed. If you blow the directory cache away entirely, you generally
2792 haven't lost any information as long as you have the name of the tree
2795 At the same time, the index is at the same time also the
2796 staging area for creating new trees, and creating a new tree always
2797 involves a controlled modification of the index file. In particular,
2798 the index file can have the representation of an intermediate tree that
2799 has not yet been instantiated. So the index can be thought of as a
2800 write-back cache, which can contain dirty information that has not yet
2801 been written back to the backing store.
2809 Generally, all "git" operations work on the index file. Some operations
2810 work *purely* on the index file (showing the current state of the
2811 index), but most operations move data to and from the index file. Either
2812 from the database or from the working directory. Thus there are four
2815 [[working-directory-to-index]]
2816 working directory -> index
2817 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2819 You update the index with information from the working directory with
2820 the gitlink:git-update-index[1] command. You
2821 generally update the index information by just specifying the filename
2822 you want to update, like so:
2824 -------------------------------------------------
2825 $ git-update-index filename
2826 -------------------------------------------------
2828 but to avoid common mistakes with filename globbing etc, the command
2829 will not normally add totally new entries or remove old entries,
2830 i.e. it will normally just update existing cache entries.
2832 To tell git that yes, you really do realize that certain files no
2833 longer exist, or that new files should be added, you
2834 should use the `--remove` and `--add` flags respectively.
2836 NOTE! A `--remove` flag does 'not' mean that subsequent filenames will
2837 necessarily be removed: if the files still exist in your directory
2838 structure, the index will be updated with their new status, not
2839 removed. The only thing `--remove` means is that update-cache will be
2840 considering a removed file to be a valid thing, and if the file really
2841 does not exist any more, it will update the index accordingly.
2843 As a special case, you can also do `git-update-index --refresh`, which
2844 will refresh the "stat" information of each index to match the current
2845 stat information. It will 'not' update the object status itself, and
2846 it will only update the fields that are used to quickly test whether
2847 an object still matches its old backing store object.
2849 [[index-to-object-database]]
2850 index -> object database
2851 ~~~~~~~~~~~~~~~~~~~~~~~~
2853 You write your current index file to a "tree" object with the program
2855 -------------------------------------------------
2857 -------------------------------------------------
2859 that doesn't come with any options - it will just write out the
2860 current index into the set of tree objects that describe that state,
2861 and it will return the name of the resulting top-level tree. You can
2862 use that tree to re-generate the index at any time by going in the
2865 [[object-database-to-index]]
2866 object database -> index
2867 ~~~~~~~~~~~~~~~~~~~~~~~~
2869 You read a "tree" file from the object database, and use that to
2870 populate (and overwrite - don't do this if your index contains any
2871 unsaved state that you might want to restore later!) your current
2872 index. Normal operation is just
2874 -------------------------------------------------
2875 $ git-read-tree <sha1 of tree>
2876 -------------------------------------------------
2878 and your index file will now be equivalent to the tree that you saved
2879 earlier. However, that is only your 'index' file: your working
2880 directory contents have not been modified.
2882 [[index-to-working-directory]]
2883 index -> working directory
2884 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2886 You update your working directory from the index by "checking out"
2887 files. This is not a very common operation, since normally you'd just
2888 keep your files updated, and rather than write to your working
2889 directory, you'd tell the index files about the changes in your
2890 working directory (i.e. `git-update-index`).
2892 However, if you decide to jump to a new version, or check out somebody
2893 else's version, or just restore a previous tree, you'd populate your
2894 index file with read-tree, and then you need to check out the result
2897 -------------------------------------------------
2898 $ git-checkout-index filename
2899 -------------------------------------------------
2901 or, if you want to check out all of the index, use `-a`.
2903 NOTE! git-checkout-index normally refuses to overwrite old files, so
2904 if you have an old version of the tree already checked out, you will
2905 need to use the "-f" flag ('before' the "-a" flag or the filename) to
2906 'force' the checkout.
2909 Finally, there are a few odds and ends which are not purely moving
2910 from one representation to the other:
2912 [[tying-it-all-together]]
2913 Tying it all together
2914 ~~~~~~~~~~~~~~~~~~~~~
2916 To commit a tree you have instantiated with "git-write-tree", you'd
2917 create a "commit" object that refers to that tree and the history
2918 behind it - most notably the "parent" commits that preceded it in
2921 Normally a "commit" has one parent: the previous state of the tree
2922 before a certain change was made. However, sometimes it can have two
2923 or more parent commits, in which case we call it a "merge", due to the
2924 fact that such a commit brings together ("merges") two or more
2925 previous states represented by other commits.
2927 In other words, while a "tree" represents a particular directory state
2928 of a working directory, a "commit" represents that state in "time",
2929 and explains how we got there.
2931 You create a commit object by giving it the tree that describes the
2932 state at the time of the commit, and a list of parents:
2934 -------------------------------------------------
2935 $ git-commit-tree <tree> -p <parent> [-p <parent2> ..]
2936 -------------------------------------------------
2938 and then giving the reason for the commit on stdin (either through
2939 redirection from a pipe or file, or by just typing it at the tty).
2941 git-commit-tree will return the name of the object that represents
2942 that commit, and you should save it away for later use. Normally,
2943 you'd commit a new `HEAD` state, and while git doesn't care where you
2944 save the note about that state, in practice we tend to just write the
2945 result to the file pointed at by `.git/HEAD`, so that we can always see
2946 what the last committed state was.
2948 Here is an ASCII art by Jon Loeliger that illustrates how
2949 various pieces fit together.
2977 checkout-index -u | | checkout-index
2988 [[examining-the-data]]
2992 You can examine the data represented in the object database and the
2993 index with various helper tools. For every object, you can use
2994 gitlink:git-cat-file[1] to examine details about the
2997 -------------------------------------------------
2998 $ git-cat-file -t <objectname>
2999 -------------------------------------------------
3001 shows the type of the object, and once you have the type (which is
3002 usually implicit in where you find the object), you can use
3004 -------------------------------------------------
3005 $ git-cat-file blob|tree|commit|tag <objectname>
3006 -------------------------------------------------
3008 to show its contents. NOTE! Trees have binary content, and as a result
3009 there is a special helper for showing that content, called
3010 `git-ls-tree`, which turns the binary content into a more easily
3013 It's especially instructive to look at "commit" objects, since those
3014 tend to be small and fairly self-explanatory. In particular, if you
3015 follow the convention of having the top commit name in `.git/HEAD`,
3018 -------------------------------------------------
3019 $ git-cat-file commit HEAD
3020 -------------------------------------------------
3022 to see what the top commit was.
3024 [[merging-multiple-trees]]
3025 Merging multiple trees
3026 ----------------------
3028 Git helps you do a three-way merge, which you can expand to n-way by
3029 repeating the merge procedure arbitrary times until you finally
3030 "commit" the state. The normal situation is that you'd only do one
3031 three-way merge (two parents), and commit it, but if you like to, you
3032 can do multiple parents in one go.
3034 To do a three-way merge, you need the two sets of "commit" objects
3035 that you want to merge, use those to find the closest common parent (a
3036 third "commit" object), and then use those commit objects to find the
3037 state of the directory ("tree" object) at these points.
3039 To get the "base" for the merge, you first look up the common parent
3042 -------------------------------------------------
3043 $ git-merge-base <commit1> <commit2>
3044 -------------------------------------------------
3046 which will return you the commit they are both based on. You should
3047 now look up the "tree" objects of those commits, which you can easily
3048 do with (for example)
3050 -------------------------------------------------
3051 $ git-cat-file commit <commitname> | head -1
3052 -------------------------------------------------
3054 since the tree object information is always the first line in a commit
3057 Once you know the three trees you are going to merge (the one "original"
3058 tree, aka the common tree, and the two "result" trees, aka the branches
3059 you want to merge), you do a "merge" read into the index. This will
3060 complain if it has to throw away your old index contents, so you should
3061 make sure that you've committed those - in fact you would normally
3062 always do a merge against your last commit (which should thus match what
3063 you have in your current index anyway).
3067 -------------------------------------------------
3068 $ git-read-tree -m -u <origtree> <yourtree> <targettree>
3069 -------------------------------------------------
3071 which will do all trivial merge operations for you directly in the
3072 index file, and you can just write the result out with
3076 [[merging-multiple-trees-2]]
3077 Merging multiple trees, continued
3078 ---------------------------------
3080 Sadly, many merges aren't trivial. If there are files that have
3081 been added.moved or removed, or if both branches have modified the
3082 same file, you will be left with an index tree that contains "merge
3083 entries" in it. Such an index tree can 'NOT' be written out to a tree
3084 object, and you will have to resolve any such merge clashes using
3085 other tools before you can write out the result.
3087 You can examine such index state with `git-ls-files --unmerged`
3088 command. An example:
3090 ------------------------------------------------
3091 $ git-read-tree -m $orig HEAD $target
3092 $ git-ls-files --unmerged
3093 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello.c
3094 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello.c
3095 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello.c
3096 ------------------------------------------------
3098 Each line of the `git-ls-files --unmerged` output begins with
3099 the blob mode bits, blob SHA1, 'stage number', and the
3100 filename. The 'stage number' is git's way to say which tree it
3101 came from: stage 1 corresponds to `$orig` tree, stage 2 `HEAD`
3102 tree, and stage3 `$target` tree.
3104 Earlier we said that trivial merges are done inside
3105 `git-read-tree -m`. For example, if the file did not change
3106 from `$orig` to `HEAD` nor `$target`, or if the file changed
3107 from `$orig` to `HEAD` and `$orig` to `$target` the same way,
3108 obviously the final outcome is what is in `HEAD`. What the
3109 above example shows is that file `hello.c` was changed from
3110 `$orig` to `HEAD` and `$orig` to `$target` in a different way.
3111 You could resolve this by running your favorite 3-way merge
3112 program, e.g. `diff3`, `merge`, or git's own merge-file, on
3113 the blob objects from these three stages yourself, like this:
3115 ------------------------------------------------
3116 $ git-cat-file blob 263414f... >hello.c~1
3117 $ git-cat-file blob 06fa6a2... >hello.c~2
3118 $ git-cat-file blob cc44c73... >hello.c~3
3119 $ git merge-file hello.c~2 hello.c~1 hello.c~3
3120 ------------------------------------------------
3122 This would leave the merge result in `hello.c~2` file, along
3123 with conflict markers if there are conflicts. After verifying
3124 the merge result makes sense, you can tell git what the final
3125 merge result for this file is by:
3127 -------------------------------------------------
3128 $ mv -f hello.c~2 hello.c
3129 $ git-update-index hello.c
3130 -------------------------------------------------
3132 When a path is in unmerged state, running `git-update-index` for
3133 that path tells git to mark the path resolved.
3135 The above is the description of a git merge at the lowest level,
3136 to help you understand what conceptually happens under the hood.
3137 In practice, nobody, not even git itself, uses three `git-cat-file`
3138 for this. There is `git-merge-index` program that extracts the
3139 stages to temporary files and calls a "merge" script on it:
3141 -------------------------------------------------
3142 $ git-merge-index git-merge-one-file hello.c
3143 -------------------------------------------------
3145 and that is what higher level `git merge -s resolve` is implemented with.
3148 How git stores objects efficiently: pack files
3149 ----------------------------------------------
3151 We've seen how git stores each object in a file named after the
3154 Unfortunately this system becomes inefficient once a project has a
3155 lot of objects. Try this on an old project:
3157 ------------------------------------------------
3159 6930 objects, 47620 kilobytes
3160 ------------------------------------------------
3162 The first number is the number of objects which are kept in
3163 individual files. The second is the amount of space taken up by
3164 those "loose" objects.
3166 You can save space and make git faster by moving these loose objects in
3167 to a "pack file", which stores a group of objects in an efficient
3168 compressed format; the details of how pack files are formatted can be
3169 found in link:technical/pack-format.txt[technical/pack-format.txt].
3171 To put the loose objects into a pack, just run git repack:
3173 ------------------------------------------------
3176 Done counting 6020 objects.
3177 Deltifying 6020 objects.
3178 100% (6020/6020) done
3179 Writing 6020 objects.
3180 100% (6020/6020) done
3181 Total 6020, written 6020 (delta 4070), reused 0 (delta 0)
3182 Pack pack-3e54ad29d5b2e05838c75df582c65257b8d08e1c created.
3183 ------------------------------------------------
3187 ------------------------------------------------
3189 ------------------------------------------------
3191 to remove any of the "loose" objects that are now contained in the
3192 pack. This will also remove any unreferenced objects (which may be
3193 created when, for example, you use "git reset" to remove a commit).
3194 You can verify that the loose objects are gone by looking at the
3195 .git/objects directory or by running
3197 ------------------------------------------------
3199 0 objects, 0 kilobytes
3200 ------------------------------------------------
3202 Although the object files are gone, any commands that refer to those
3203 objects will work exactly as they did before.
3205 The gitlink:git-gc[1] command performs packing, pruning, and more for
3206 you, so is normally the only high-level command you need.
3208 [[dangling-objects]]
3212 The gitlink:git-fsck[1] command will sometimes complain about dangling
3213 objects. They are not a problem.
3215 The most common cause of dangling objects is that you've rebased a
3216 branch, or you have pulled from somebody else who rebased a branch--see
3217 <<cleaning-up-history>>. In that case, the old head of the original
3218 branch still exists, as does everything it pointed to. The branch
3219 pointer itself just doesn't, since you replaced it with another one.
3221 There are also other situations that cause dangling objects. For
3222 example, a "dangling blob" may arise because you did a "git add" of a
3223 file, but then, before you actually committed it and made it part of the
3224 bigger picture, you changed something else in that file and committed
3225 that *updated* thing - the old state that you added originally ends up
3226 not being pointed to by any commit or tree, so it's now a dangling blob
3229 Similarly, when the "recursive" merge strategy runs, and finds that
3230 there are criss-cross merges and thus more than one merge base (which is
3231 fairly unusual, but it does happen), it will generate one temporary
3232 midway tree (or possibly even more, if you had lots of criss-crossing
3233 merges and more than two merge bases) as a temporary internal merge
3234 base, and again, those are real objects, but the end result will not end
3235 up pointing to them, so they end up "dangling" in your repository.
3237 Generally, dangling objects aren't anything to worry about. They can
3238 even be very useful: if you screw something up, the dangling objects can
3239 be how you recover your old tree (say, you did a rebase, and realized
3240 that you really didn't want to - you can look at what dangling objects
3241 you have, and decide to reset your head to some old dangling state).
3243 For commits, you can just use:
3245 ------------------------------------------------
3246 $ gitk <dangling-commit-sha-goes-here> --not --all
3247 ------------------------------------------------
3249 This asks for all the history reachable from the given commit but not
3250 from any branch, tag, or other reference. If you decide it's something
3251 you want, you can always create a new reference to it, e.g.,
3253 ------------------------------------------------
3254 $ git branch recovered-branch <dangling-commit-sha-goes-here>
3255 ------------------------------------------------
3257 For blobs and trees, you can't do the same, but you can still examine
3258 them. You can just do
3260 ------------------------------------------------
3261 $ git show <dangling-blob/tree-sha-goes-here>
3262 ------------------------------------------------
3264 to show what the contents of the blob were (or, for a tree, basically
3265 what the "ls" for that directory was), and that may give you some idea
3266 of what the operation was that left that dangling object.
3268 Usually, dangling blobs and trees aren't very interesting. They're
3269 almost always the result of either being a half-way mergebase (the blob
3270 will often even have the conflict markers from a merge in it, if you
3271 have had conflicting merges that you fixed up by hand), or simply
3272 because you interrupted a "git fetch" with ^C or something like that,
3273 leaving _some_ of the new objects in the object database, but just
3274 dangling and useless.
3276 Anyway, once you are sure that you're not interested in any dangling
3277 state, you can just prune all unreachable objects:
3279 ------------------------------------------------
3281 ------------------------------------------------
3283 and they'll be gone. But you should only run "git prune" on a quiescent
3284 repository - it's kind of like doing a filesystem fsck recovery: you
3285 don't want to do that while the filesystem is mounted.
3287 (The same is true of "git-fsck" itself, btw - but since
3288 git-fsck never actually *changes* the repository, it just reports
3289 on what it found, git-fsck itself is never "dangerous" to run.
3290 Running it while somebody is actually changing the repository can cause
3291 confusing and scary messages, but it won't actually do anything bad. In
3292 contrast, running "git prune" while somebody is actively changing the
3293 repository is a *BAD* idea).
3295 [[birdview-on-the-source-code]]
3296 A birds-eye view of Git's source code
3297 -------------------------------------
3299 It is not always easy for new developers to find their way through Git's
3300 source code. This section gives you a little guidance to show where to
3303 A good place to start is with the contents of the initial commit, with:
3305 ----------------------------------------------------
3306 $ git checkout e83c5163
3307 ----------------------------------------------------
3309 The initial revision lays the foundation for almost everything git has
3310 today, but is small enough to read in one sitting.
3312 Note that terminology has changed since that revision. For example, the
3313 README in that revision uses the word "changeset" to describe what we
3314 now call a <<def_commit_object,commit>>.
3316 Also, we do not call it "cache" any more, but "index", however, the
3317 file is still called `cache.h`. Remark: Not much reason to change it now,
3318 especially since there is no good single name for it anyway, because it is
3319 basically _the_ header file which is included by _all_ of Git's C sources.
3321 If you grasp the ideas in that initial commit, you should check out a
3322 more recent version and skim `cache.h`, `object.h` and `commit.h`.
3324 In the early days, Git (in the tradition of UNIX) was a bunch of programs
3325 which were extremely simple, and which you used in scripts, piping the
3326 output of one into another. This turned out to be good for initial
3327 development, since it was easier to test new things. However, recently
3328 many of these parts have become builtins, and some of the core has been
3329 "libified", i.e. put into libgit.a for performance, portability reasons,
3330 and to avoid code duplication.
3332 By now, you know what the index is (and find the corresponding data
3333 structures in `cache.h`), and that there are just a couple of object types
3334 (blobs, trees, commits and tags) which inherit their common structure from
3335 `struct object`, which is their first member (and thus, you can cast e.g.
3336 `(struct object *)commit` to achieve the _same_ as `&commit->object`, i.e.
3337 get at the object name and flags).
3339 Now is a good point to take a break to let this information sink in.
3341 Next step: get familiar with the object naming. Read <<naming-commits>>.
3342 There are quite a few ways to name an object (and not only revisions!).
3343 All of these are handled in `sha1_name.c`. Just have a quick look at
3344 the function `get_sha1()`. A lot of the special handling is done by
3345 functions like `get_sha1_basic()` or the likes.
3347 This is just to get you into the groove for the most libified part of Git:
3348 the revision walker.
3350 Basically, the initial version of `git log` was a shell script:
3352 ----------------------------------------------------------------
3353 $ git-rev-list --pretty $(git-rev-parse --default HEAD "$@") | \
3354 LESS=-S ${PAGER:-less}
3355 ----------------------------------------------------------------
3357 What does this mean?
3359 `git-rev-list` is the original version of the revision walker, which
3360 _always_ printed a list of revisions to stdout. It is still functional,
3361 and needs to, since most new Git programs start out as scripts using
3364 `git-rev-parse` is not as important any more; it was only used to filter out
3365 options that were relevant for the different plumbing commands that were
3366 called by the script.
3368 Most of what `git-rev-list` did is contained in `revision.c` and
3369 `revision.h`. It wraps the options in a struct named `rev_info`, which
3370 controls how and what revisions are walked, and more.
3372 The original job of `git-rev-parse` is now taken by the function
3373 `setup_revisions()`, which parses the revisions and the common command line
3374 options for the revision walker. This information is stored in the struct
3375 `rev_info` for later consumption. You can do your own command line option
3376 parsing after calling `setup_revisions()`. After that, you have to call
3377 `prepare_revision_walk()` for initialization, and then you can get the
3378 commits one by one with the function `get_revision()`.
3380 If you are interested in more details of the revision walking process,
3381 just have a look at the first implementation of `cmd_log()`; call
3382 `git-show v1.3.0~155^2~4` and scroll down to that function (note that you
3383 no longer need to call `setup_pager()` directly).
3385 Nowadays, `git log` is a builtin, which means that it is _contained_ in the
3386 command `git`. The source side of a builtin is
3388 - a function called `cmd_<bla>`, typically defined in `builtin-<bla>.c`,
3389 and declared in `builtin.h`,
3391 - an entry in the `commands[]` array in `git.c`, and
3393 - an entry in `BUILTIN_OBJECTS` in the `Makefile`.
3395 Sometimes, more than one builtin is contained in one source file. For
3396 example, `cmd_whatchanged()` and `cmd_log()` both reside in `builtin-log.c`,
3397 since they share quite a bit of code. In that case, the commands which are
3398 _not_ named like the `.c` file in which they live have to be listed in
3399 `BUILT_INS` in the `Makefile`.
3401 `git log` looks more complicated in C than it does in the original script,
3402 but that allows for a much greater flexibility and performance.
3404 Here again it is a good point to take a pause.
3406 Lesson three is: study the code. Really, it is the best way to learn about
3407 the organization of Git (after you know the basic concepts).
3409 So, think about something which you are interested in, say, "how can I
3410 access a blob just knowing the object name of it?". The first step is to
3411 find a Git command with which you can do it. In this example, it is either
3412 `git show` or `git cat-file`.
3414 For the sake of clarity, let's stay with `git cat-file`, because it
3418 - was around even in the initial commit (it literally went only through
3419 some 20 revisions as `cat-file.c`, was renamed to `builtin-cat-file.c`
3420 when made a builtin, and then saw less than 10 versions).
3422 So, look into `builtin-cat-file.c`, search for `cmd_cat_file()` and look what
3425 ------------------------------------------------------------------
3426 git_config(git_default_config);
3428 usage("git-cat-file [-t|-s|-e|-p|<type>] <sha1>");
3429 if (get_sha1(argv[2], sha1))
3430 die("Not a valid object name %s", argv[2]);
3431 ------------------------------------------------------------------
3433 Let's skip over the obvious details; the only really interesting part
3434 here is the call to `get_sha1()`. It tries to interpret `argv[2]` as an
3435 object name, and if it refers to an object which is present in the current
3436 repository, it writes the resulting SHA-1 into the variable `sha1`.
3438 Two things are interesting here:
3440 - `get_sha1()` returns 0 on _success_. This might surprise some new
3441 Git hackers, but there is a long tradition in UNIX to return different
3442 negative numbers in case of different errors -- and 0 on success.
3444 - the variable `sha1` in the function signature of `get_sha1()` is `unsigned
3445 char \*`, but is actually expected to be a pointer to `unsigned
3446 char[20]`. This variable will contain the 160-bit SHA-1 of the given
3447 commit. Note that whenever a SHA-1 is passed as `unsigned char \*`, it
3448 is the binary representation, as opposed to the ASCII representation in
3449 hex characters, which is passed as `char *`.
3451 You will see both of these things throughout the code.
3455 -----------------------------------------------------------------------------
3457 buf = read_object_with_reference(sha1, argv[1], &size, NULL);
3458 -----------------------------------------------------------------------------
3460 This is how you read a blob (actually, not only a blob, but any type of
3461 object). To know how the function `read_object_with_reference()` actually
3462 works, find the source code for it (something like `git grep
3463 read_object_with | grep ":[a-z]"` in the git repository), and read
3466 To find out how the result can be used, just read on in `cmd_cat_file()`:
3468 -----------------------------------
3469 write_or_die(1, buf, size);
3470 -----------------------------------
3472 Sometimes, you do not know where to look for a feature. In many such cases,
3473 it helps to search through the output of `git log`, and then `git show` the
3474 corresponding commit.
3476 Example: If you know that there was some test case for `git bundle`, but
3477 do not remember where it was (yes, you _could_ `git grep bundle t/`, but that
3478 does not illustrate the point!):
3480 ------------------------
3481 $ git log --no-merges t/
3482 ------------------------
3484 In the pager (`less`), just search for "bundle", go a few lines back,
3485 and see that it is in commit 18449ab0... Now just copy this object name,
3486 and paste it into the command line
3494 Another example: Find out what to do in order to make some script a
3497 -------------------------------------------------
3498 $ git log --no-merges --diff-filter=A builtin-*.c
3499 -------------------------------------------------
3501 You see, Git is actually the best tool to find out about the source of Git
3505 include::glossary.txt[]
3508 Appendix A: Git Quick Start
3509 ===========================
3511 This is a quick summary of the major commands; the following chapters
3512 will explain how these work in more detail.
3514 [[quick-creating-a-new-repository]]
3515 Creating a new repository
3516 -------------------------
3520 -----------------------------------------------
3521 $ tar xzf project.tar.gz
3524 Initialized empty Git repository in .git/
3527 -----------------------------------------------
3529 From a remote repository:
3531 -----------------------------------------------
3532 $ git clone git://example.com/pub/project.git
3534 -----------------------------------------------
3536 [[managing-branches]]
3540 -----------------------------------------------
3541 $ git branch # list all local branches in this repo
3542 $ git checkout test # switch working directory to branch "test"
3543 $ git branch new # create branch "new" starting at current HEAD
3544 $ git branch -d new # delete branch "new"
3545 -----------------------------------------------
3547 Instead of basing new branch on current HEAD (the default), use:
3549 -----------------------------------------------
3550 $ git branch new test # branch named "test"
3551 $ git branch new v2.6.15 # tag named v2.6.15
3552 $ git branch new HEAD^ # commit before the most recent
3553 $ git branch new HEAD^^ # commit before that
3554 $ git branch new test~10 # ten commits before tip of branch "test"
3555 -----------------------------------------------
3557 Create and switch to a new branch at the same time:
3559 -----------------------------------------------
3560 $ git checkout -b new v2.6.15
3561 -----------------------------------------------
3563 Update and examine branches from the repository you cloned from:
3565 -----------------------------------------------
3566 $ git fetch # update
3567 $ git branch -r # list
3571 $ git checkout -b masterwork origin/master
3572 -----------------------------------------------
3574 Fetch a branch from a different repository, and give it a new
3575 name in your repository:
3577 -----------------------------------------------
3578 $ git fetch git://example.com/project.git theirbranch:mybranch
3579 $ git fetch git://example.com/project.git v2.6.15:mybranch
3580 -----------------------------------------------
3582 Keep a list of repositories you work with regularly:
3584 -----------------------------------------------
3585 $ git remote add example git://example.com/project.git
3586 $ git remote # list remote repositories
3589 $ git remote show example # get details
3591 URL: git://example.com/project.git
3592 Tracked remote branches
3594 $ git fetch example # update branches from example
3595 $ git branch -r # list all remote branches
3596 -----------------------------------------------
3599 [[exploring-history]]
3603 -----------------------------------------------
3604 $ gitk # visualize and browse history
3605 $ git log # list all commits
3606 $ git log src/ # ...modifying src/
3607 $ git log v2.6.15..v2.6.16 # ...in v2.6.16, not in v2.6.15
3608 $ git log master..test # ...in branch test, not in branch master
3609 $ git log test..master # ...in branch master, but not in test
3610 $ git log test...master # ...in one branch, not in both
3611 $ git log -S'foo()' # ...where difference contain "foo()"
3612 $ git log --since="2 weeks ago"
3613 $ git log -p # show patches as well
3614 $ git show # most recent commit
3615 $ git diff v2.6.15..v2.6.16 # diff between two tagged versions
3616 $ git diff v2.6.15..HEAD # diff with current head
3617 $ git grep "foo()" # search working directory for "foo()"
3618 $ git grep v2.6.15 "foo()" # search old tree for "foo()"
3619 $ git show v2.6.15:a.txt # look at old version of a.txt
3620 -----------------------------------------------
3622 Search for regressions:
3624 -----------------------------------------------
3626 $ git bisect bad # current version is bad
3627 $ git bisect good v2.6.13-rc2 # last known good revision
3628 Bisecting: 675 revisions left to test after this
3630 $ git bisect good # if this revision is good, or
3631 $ git bisect bad # if this revision is bad.
3632 # repeat until done.
3633 -----------------------------------------------
3639 Make sure git knows who to blame:
3641 ------------------------------------------------
3642 $ cat >>~/.gitconfig <<\EOF
3644 name = Your Name Comes Here
3645 email = you@yourdomain.example.com
3647 ------------------------------------------------
3649 Select file contents to include in the next commit, then make the
3652 -----------------------------------------------
3653 $ git add a.txt # updated file
3654 $ git add b.txt # new file
3655 $ git rm c.txt # old file
3657 -----------------------------------------------
3659 Or, prepare and create the commit in one step:
3661 -----------------------------------------------
3662 $ git commit d.txt # use latest content only of d.txt
3663 $ git commit -a # use latest content of all tracked files
3664 -----------------------------------------------
3670 -----------------------------------------------
3671 $ git merge test # merge branch "test" into the current branch
3672 $ git pull git://example.com/project.git master
3673 # fetch and merge in remote branch
3674 $ git pull . test # equivalent to git merge test
3675 -----------------------------------------------
3677 [[sharing-your-changes]]
3678 Sharing your changes
3679 --------------------
3681 Importing or exporting patches:
3683 -----------------------------------------------
3684 $ git format-patch origin..HEAD # format a patch for each commit
3685 # in HEAD but not in origin
3686 $ git am mbox # import patches from the mailbox "mbox"
3687 -----------------------------------------------
3689 Fetch a branch in a different git repository, then merge into the
3692 -----------------------------------------------
3693 $ git pull git://example.com/project.git theirbranch
3694 -----------------------------------------------
3696 Store the fetched branch into a local branch before merging into the
3699 -----------------------------------------------
3700 $ git pull git://example.com/project.git theirbranch:mybranch
3701 -----------------------------------------------
3703 After creating commits on a local branch, update the remote
3704 branch with your commits:
3706 -----------------------------------------------
3707 $ git push ssh://example.com/project.git mybranch:theirbranch
3708 -----------------------------------------------
3710 When remote and local branch are both named "test":
3712 -----------------------------------------------
3713 $ git push ssh://example.com/project.git test
3714 -----------------------------------------------
3716 Shortcut version for a frequently used remote repository:
3718 -----------------------------------------------
3719 $ git remote add example ssh://example.com/project.git
3720 $ git push example test
3721 -----------------------------------------------
3723 [[repository-maintenance]]
3724 Repository maintenance
3725 ----------------------
3727 Check for corruption:
3729 -----------------------------------------------
3731 -----------------------------------------------
3733 Recompress, remove unused cruft:
3735 -----------------------------------------------
3737 -----------------------------------------------
3741 Appendix B: Notes and todo list for this manual
3742 ===============================================
3744 This is a work in progress.
3746 The basic requirements:
3747 - It must be readable in order, from beginning to end, by
3748 someone intelligent with a basic grasp of the unix
3749 commandline, but without any special knowledge of git. If
3750 necessary, any other prerequisites should be specifically
3751 mentioned as they arise.
3752 - Whenever possible, section headings should clearly describe
3753 the task they explain how to do, in language that requires
3754 no more knowledge than necessary: for example, "importing
3755 patches into a project" rather than "the git-am command"
3757 Think about how to create a clear chapter dependency graph that will
3758 allow people to get to important topics without necessarily reading
3759 everything in between.
3761 Say something about .gitignore.
3763 Scan Documentation/ for other stuff left out; in particular:
3767 list of commands in gitlink:git[1]
3769 Scan email archives for other stuff left out
3771 Scan man pages to see if any assume more background than this manual
3774 Simplify beginning by suggesting disconnected head instead of
3775 temporary branch creation?
3777 Add more good examples. Entire sections of just cookbook examples
3778 might be a good idea; maybe make an "advanced examples" section a
3779 standard end-of-chapter section?
3781 Include cross-references to the glossary, where appropriate.
3783 Document shallow clones? See draft 1.5.0 release notes for some
3786 Add a section on working with other version control systems, including
3787 CVS, Subversion, and just imports of series of release tarballs.
3789 More details on gitweb?
3791 Write a chapter on using plumbing and writing scripts.